fusion809 has quit [Remote host closed the connection]
<nixtacular>
how do i override the attributes in the sets that correspond to `linuxPackages_4_19`, `linuxPackages_4_14`, etc.? i'm trying to use overlays for this purpose and it's causing Nix to not see whichever `linuxPackages_X_Y` set i'm overriding
<nixtacular>
i have a hunch that it has something to do with the fact that overlays are meant for derivations and `linuxPackages_X_Y` isn't a derivation, it's a set
<nixtacular>
but then i don't know what to do, short of checking out my own version of nixpkgs :-/
<gchristensen>
oh man I fought this same battle and figured it out once
<ajs124>
On a completly unrelated note: can subversion do something like git clone --depth=1 ?
<nixtacular>
@ajs124, that's what i'm trying to do right now, but I get the error `error: selector 'linuxPackages_4_19.prl-tools' matches no derivations`
<gchristensen>
subversion doesn't fetch all the history anyway
<gchristensen>
ajs124: so, no need to --depth=1 it
redd has joined #nixos
<nixtacular>
gchristensen: bummer :-/
redd is now known as Guest90927
<gchristensen>
nixtacular: bummer?
<nixtacular>
gchristensen: American English saying for "that's unfortunate" :-)
Guest90927 is now known as redd_using_nixos
<gchristensen>
totally, but what is a bummer?
<spinus>
ryantm: thank you, filterSource - I had somewhere else other package which swamped all repo by accident
<nixtacular>
oh sorry. just that you weren't able to find the solution. i've been banging my head against the wall for 3 hours trying to figure this out lol
alex has quit [Ping timeout: 260 seconds]
alexherbo2 has quit [Ping timeout: 268 seconds]
<gchristensen>
ahh
<gchristensen>
it looks like prl-tools was dropped in kernels after 4.14
<nixtacular>
gchristensen: correct. i'm attempting to fix the package.
<gchristensen>
ah
<ajs124>
gchristensen: if svn is shallow anyways, why does it create a bunch of subdirectories in tags/ and generally seem to downloads hundreds of MBs if one tag is ~10MB?
<nixtacular>
`linuxPackages.prl-tools` basically gets set to `null` for current versions of NixOS
andersk_ has quit [Remote host closed the connection]
redd_using_nixos has quit [Remote host closed the connection]
<ajs124>
I think I got it. subversion is confusing -.-
vidbina has quit [Ping timeout: 268 seconds]
pie_ has quit [Quit: pie_]
<gchristensen>
cvs, now that is confusing
mexisme has joined #nixos
<gchristensen>
hope you never decide to rename or delete a directory!
zeta_0 has joined #nixos
ng0 has quit [Quit: leaving]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JexCM
<nixtacular>
ajs124: you linked to a gist that you said works for overriding `linuxPackages`. could you give more context as to where it's used?
<nixtacular>
i'm trying to use that same basic approach in an overlay but it doesn't work
<nixtacular>
i'm also trying to figure out exactly how/where overlays get applied but i can't find the file
Rusty1 has quit [Remote host closed the connection]
thc202 has quit [Ping timeout: 248 seconds]
<zeta_0>
in haskell, normally for a yesod language dsl i would need to place `{-# LanguageExtension #-}` on the top of the haskell file or add the language extension through the `ghc command line` by passing an extra argument, or in a cabal file add a `default-extensions block`, how do i add this to a default.nix file instead?
<{^_^}>
[nixpkgs] @pasqui23 opened pull request #76805 → nodePackages.dat:added dependencies to fix compilation → https://git.io/JexCF
<zeta_0>
clever: can i add haskell language pragmas to default.nix, it would save me the hassle of having to manually type them out over and over again?
<clever>
zeta_0: dont know, but haskell.nix should obey the cabal file
Rusty1 has joined #nixos
ng0 has joined #nixos
bvdw has quit [Read error: Connection reset by peer]
ng0 has quit [Client Quit]
bvdw has joined #nixos
orivej has joined #nixos
<zeta_0>
clever: well, i don't know if i can do it with a default.nix file, so instead, in the cabal file i will add a default-extensions block like the yesod book says, then do cabal2nix, this should save me from having to manually rewrite the language pragramas over and over again, which would be very painful
ng0 has joined #nixos
ng0 has joined #nixos
nixtacular has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 240 seconds]
andersk has joined #nixos
<zeta_0>
clever: in the elisp terminal when i enter nix-shell, it loses the auto-completion for all the commands, and it also loses all of my .emacs configuration, is there a way to add this stuff to nix-shell? should i ask in the #emacs channel instead?
pie_ has joined #nixos
<clever>
zeta_0: probably an emacs issue
pie_ has quit [Client Quit]
<zeta_0>
clever: ok, i asked this in #emacs instead
<lovesegfault>
eoli3n_: Had to go out with the wife, yeah mocin i3->sway is a big change (not because of the WM but the ancillary tools)
<colemickens>
but it had to donwload hundreds of megs of deps in the process.
ng0 has quit [Quit: leaving]
<colemickens>
preferably I could say nix-build --only-build-if-no-subs-exist-on-any-cache and then it would just no-op because it knows all closures have been built because it finds them on my cachix cache.
<gchristensen>
so you don't want the build result?
<colemickens>
I can implement this by saying "am I only a CI build, are there no git changes from my update script, skip building" but that feels more brittle than letting Nix figure it out.
<colemickens>
In this case, nah, I don't need the build result actually available at the end.
h0m1 has quit [Ping timeout: 260 seconds]
<gchristensen>
if `nix-build` with `-j0` exits 0, it means no builds are necessariy
<gchristensen>
I believe that would work, maybe test that
<colemickens>
!
<gchristensen>
it may still substitute though ...
zeta_0 has quit [Quit: taking break]
<gchristensen>
colemickens: you could manually query your configured caches, though that is a bit weird: /nix/store/00zf0sh2yxsmi083my3daaxc8iy5ynga-curl-7.67.0-dev => cache.nixos.org/00zf0sh2yxsmi083my3daaxc8iy5ynga.narinfo => 200 OK => already cached
h0m1 has joined #nixos
selfsymmetric-mu has joined #nixos
<colemickens>
I can get nix-instantiate to just spit out hte drvs. I can open the derivation and see the output path, but I don't know how to get that programmatically easily. `nix path-info --json ` on the deriv doesn't contain the output path.
<shajra>
Is there already an API in Nixpkgs that makes patching source as simple as filtering it?
zeta_0 has joined #nixos
<colemickens>
If I could do that, nix-instantiate -> loop over drvs -> check if path exists on my cache, could be good enough, but might not be any cleaner/happier than just checking if git is dirty
<zeta_0>
it says there is a manual available: nix-mode.org but that link isn't working?
<gchristensen>
colemickens: you only need to check the top level drvs
<shajra>
I know someone will mention the "preConfigure" and such hooks of the derivation itself, but this is for Haskell.nix, and I want to much source before it gets to the derivation. Because Haskell.nix parses it first to make a compilation plan.
<gchristensen>
clever: I think you're pretty familiar with haskell.nix? shajra needs a hand
<colemickens>
Is that because the drv is only pushed if the build completes? Is that assumption invalidated if I run cachix in store watch mode?
<gchristensen>
colemickens: sorry, I mean you only need to check the output path of the top level drv
<clever>
shajra: if you want to patch the source, you probably want either patchPhase, or use a custom runCommand based thing, to compose several things together, and use the result as a src
<gchristensen>
colemickens: and it is true because it is an invariant that if a path exists in a cache, all of its (run-time) dependencies are in the cache as well
<shajra>
gchristensen: thanks for the help getting help... I think you saw the previous problem on Twitter... man... that was frustrating, but I'm past that.
<gchristensen>
shajra: :)
<colemickens>
gchristensen: understood, I just misunderstood the "top-level" as meaning drv, which was silly to have assumed. I think I'm on the same page.
<selfsymmetric-mu>
zeta_0: https://github.com/NixOS/nix-mode It has nix-shell and nix-repl wrappers, but I haven't investigated them deeply.
<shajra>
clever: yeah, if you know Haskell.nix, you probably know what I'm dealing with. It's making a plan before it's even getting to the derivation.
<shajra>
clever: in fact, that's exactly what I'm trying to do... patch cabal.project so I get a different plan. cabal.project isn't even put into the workspace for the derivation.
<shajra>
clever: but if runCommand is the way to go, I can do that. Sometimes these simple things end up more verbose than I feel they should be.
<{^_^}>
selfsymmetric-mu's karma got increased to 3
<clever>
shajra: i dont think cabal.project should exist at build time, haskell.nix is supposed to parse cabal.project, then build everything with old-cabal
<clever>
i think
<shajra>
clever: yeah... I want to add in "source-repository-package" entry, just for my Haskell.nix build... because... well... (doctest).
<clever>
shajra: havent tried doing anything like that yet
<shajra>
clever: yeah, I'm pretty sure I'm just one patched src away from having this working.
<wucke13>
How to debug `cp: cannot create directory 'source-/<hash>-source'` on a multi src expression?
<clever>
wucke13: when cp copies things from the store, the copy it creates will be read-only, you may need a chmod +w -R
<wucke13>
In what phase?
<shajra>
clever: I think in general, I don't want any fancy non-standard building for this project, because it would be nice for it to end up on Hackage, and as such it should build with what's on Hackage... like the latest `cabal-doctest` package there. But unfortunately, because Haskell.nix builds a component at a time by calling `Setup.hs` directly, it doesn't jive well with `cabal-doctest` as it exists now.
<clever>
wucke13: between creating the destination, and trying to copy a 2nd thing to it
<clever>
wucke13: note, that you can never write to $src, only $out and subdirs of the temp dir you start in
<shajra>
clever: I have a PR up, but in the meantime, I'm making my Haskell.nix build use a fork of `cabal-doctest` which seems to work.
silver has quit [Read error: Connection reset by peer]
<{^_^}>
[cabal2nix] @andersk opened pull request #438 → Fix unsafe concatenation of $LD_LIBRARY_PATH → https://git.io/JexW7
<wucke13>
clever: Can you give me a pointer about this? I have nothing in my expression on how to unpack, just `srcs = [ src1 src2 ];`. To me it's unclear what can fail on that basic situation.
<clever>
wucke13: can you pastebin more of the nix expression?
<clever>
wucke13: buildRustPackage doesnt accept a srcs argument
<clever>
wucke13: and sourceRoot = pname; likely doesnt do what you expect it does
<wucke13>
clever: What does it do?
<clever>
wucke13: sourceRoot tells the stdenv where to cd, after the unpackPhase has unpacked things
<clever>
wucke13: and the unpackPhase is likely creating a directory called source
<clever>
so changing sourceRoot, means it will try to `cd screenruster`, which doesnt exist
<clever>
wucke13: which repo do you want to be cd'd into when the build starts? where should the other repo exist?
<wucke13>
CWD=$buildroot/screenruster. Following layout: $buildroot/{screenruster,saver}/ (yes, screenruster-saver needs to be renamed to saver too.
<clever>
wucke13: rename line 8 to just src=, delete line 23&24, and then try `postUnpack = "cp -r $screenruster-saver-src saver ; pwd ; ls -lh";`
<wucke13>
clever: Thanks for the clear instructions. Unfortunately that yields some erros from a cargo nix expression. And it looks like the unpack phase is not ran at all: https://paste.krebsco.de/wxUgmqcN
knupfer has joined #nixos
<clever>
5 , src ? null
<clever>
6 , srcs ? null
<clever>
oh, it does accept a srcs
<clever>
26 then fetchcargo {
<clever>
27 inherit name src srcs sourceRoot cargoUpdateHook;
<clever>
which it feeds directly to fetchcargo
<{^_^}>
[nixos-org-configurations] @grahamc pushed to master « status page: move the noscript bit out of the table, since that didn't work anyway »: https://git.io/Jexlc
<clever>
so fetchcargo is being ran in its own derivation, before the custom postUnpack
<clever>
wucke13: so it will run the default unpackPhase
<clever>
wucke13: line 869 will run unpackFile on each thing in $srcs
<clever>
wucke13: ahhh, i see the problem
<clever>
wucke13: when it copies j6q42xy7c9f0p9qw16i9hdy8768rys0m-source, it strips the hash, creating source/
<clever>
wucke13: then when it tries to copy fv7byzwlavjz1ad9g4zrwv0ddby78b08-source to source, it already exists, so it winds up copying to source/fv7byzwlavjz1ad9g4zrwv0ddby78b08-source
<clever>
wucke13: you need to name both of the fetchFromGitHub's
<clever>
wucke13: just add a name= on each
<wucke13>
Sweet, that seems to do at least something! Thanks for the detailed investigation so far :D
<tpham1>
How can I force patchelf set RPATH instead of RUNPATH
<tpham1>
RUNPATH is ignored on CentOS
* colemickens
is reading the tweag article about nix-build hooks and untrusted CI... how does hook ordering work?
<gchristensen>
colemickens: what do you mean hook ordering?
<colemickens>
eh, it looks like there's only one hook and it's configured by root anyway
<colemickens>
I was wondering if a user could get a malicious hook to trigger before the "trusted" sign/upload hook.
justanotheruser has quit [Ping timeout: 248 seconds]
<tomberek>
zeta_0: in initExtra or sessionVariables you should be able to modify PS1
rogue_koder_ has joined #nixos
Soo_Slow has quit [Client Quit]
<zeta_0>
tomberek: ok, so how would i set: `\u@\h:\W\$` in `sessionVariables`?
<tomberek>
zeta_0: not 100% sure the order things are applied, so seesionVariables may not work, but it would be something like `sessionVariables.PS1 = "\u@\h:\W\$"`
<tomberek>
i'll test now
<tomberek>
okay, initExtra works. that's just a raw string where you can have "export PS1=whatever"
<tomberek>
and sessionVariables works too
<tomberek>
cleaner is probably programs.bash.sessionVariables.PS1 = "whatever";
zupo has quit [Ping timeout: 268 seconds]
<zeta_0>
tomberek: programs.bash.sessionVariables.PS1 = "\u@\h:\W\$"; ,for sessionVariables and "export PS1=\u@\h:\W\$" for initExtra ?
<tomberek>
one or the other, don't need both
<zeta_0>
what is the difference between the 2?
<zeta_0>
initExtra and sessionVariable bash options?
<zeta_0>
tomberek: ^
<tomberek>
i'd have to look at the program.bash implementation, but i suspect that as you write it there will be no difference. the "initExtra" exposes raw bash you can write, while the sessionVariables provides a bit more structure.
<zeta_0>
tomberek: ok, i'll try out the sessionVariables option, i'll let you know if it works or not
<tomberek>
zeta_0: that's just exposing the bash HISTCONTROL variable, so yes. Read the bash man page to get a full description of what that does.
<tomberek>
bash only allows one of those at a time
<zeta_0>
ok, i think i will go with "erasedups" for now
<zeta_0>
time for a rebuild
<tomberek>
i think it can also be ignoreboth
<tomberek>
hm... nope.. some logic prevents that
rogue_koder_ has quit [Quit: Konversation terminated!]
zeta_0 has quit [Remote host closed the connection]
takeda has joined #nixos
zeta_0 has joined #nixos
<zeta_0>
tomberek: i was able to do a rebuild switch without any errors, but the path does not stay shortened when i go into more subdirectories?
<zeta_0>
tomberek: maybe \u@\h:\W\$ is wrong and i need to do something else?
jlv has quit [Remote host closed the connection]
<zeta_0>
sessionVariables = {
<zeta_0>
};
<zeta_0>
PS1 = "\u@\h:\W\$";
<zeta_0>
this isn't keep my bash path short guys, whats wrong?
<tomberek>
zeta_0: i don't recall exactly what those do. and by the way, you don't need to do a rebuild and reboot, "home-manager switch" and creating a new terminal is enough
<tomberek>
\u = username \W is basename of CWD with $HOME abbreviated.... what exactly do you you want?
captn3m0- has joined #nixos
<zeta_0>
tomberek: when i cd into new directories i want it to show only the current directory that i am in, like how i have my elisp shell setup in emacs:
<zeta_0>
[zeta@nixos:/nix-haskell-recipes]$ pwd
takeda has quit [Ping timeout: 260 seconds]
captn3m0 has quit [Ping timeout: 268 seconds]
<tomberek>
looks like that's what "W" does, you sure it's capitalized? plus check to make sure the effect took by `echo "$PS"`
lordcirth_ has quit [Remote host closed the connection]
felixfoertsch23 has quit [Quit: ZNC 1.7.3 - https://znc.in]
Rusty1 has quit [Quit: Konversation terminated!]
felixfoertsch has joined #nixos
takeda has joined #nixos
takeda has quit [Ping timeout: 240 seconds]
chagra has quit [Ping timeout: 260 seconds]
palo1 has joined #nixos
palo has quit [Ping timeout: 260 seconds]
palo1 is now known as palo
Soo_Slow has joined #nixos
chagra has joined #nixos
ddellacosta has joined #nixos
hatefulmoron has joined #nixos
endformationage has quit [Quit: WeeChat 2.6]
hatefulmoron has quit [Quit: WeeChat 2.6]
ddellacosta has quit [Ping timeout: 258 seconds]
takeda has joined #nixos
hatefulmoron has joined #nixos
<hatefulmoron>
Hi, I'm having some trouble installing the `weechat-matrix` plugin for weechat from the unstable channel. I've included how I've installed the package, what error I get, and how I've tried to fix it here: https://pastebin.com/LU6FuHsL
takeda has quit [Ping timeout: 258 seconds]
tomberek has quit [Ping timeout: 240 seconds]
chagra has quit [Ping timeout: 258 seconds]
Phillemann has quit [Quit: WeeChat 2.7]
PLPD-Bot has quit [Remote host closed the connection]
tpham1 has left #nixos [#nixos]
ebzzry has joined #nixos
tpham1 has joined #nixos
takeda has joined #nixos
tomberek has joined #nixos
MichaelRaskin has quit [Quit: MichaelRaskin]
xkapastel has quit [Quit: Connection closed for inactivity]
<hatefulmoron>
As an update to my situation, I have weechat running by removing all references of the `typing` module of python 3.7 from PYTHONPATH, because apparently having that module installed on python 3.7 causes issues.
<jtojnar>
Izorkin hmm, it appears that by claiming fail2ban supports python 3, they meant that you can run 2to3 of it 🤦♀️️
felixfoertsch has quit [Quit: ZNC 1.7.3 - https://znc.in]
spinus has joined #nixos
<flokli>
eoli3n_: you never ever want to set stateVersion to something else than the one you set on boot
<flokli>
,stateVersion
<{^_^}>
Setting stateVersion to the latest release doesn't upgrade anything and can only break your setup at best. To actually upgrade NixOS see https://nixos.org/nixos/manual/#sec-upgrading. If you want to update the stateVersion option regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually migrated with the new value.
takeda has joined #nixos
<eoli3n_>
ok thx
felixfoertsch has joined #nixos
takeda has quit [Ping timeout: 240 seconds]
<{^_^}>
[nix] @edolstra merged pull request #3296 → exportReferencesGraph: support working → https://git.io/JexWB
<eoli3n_>
in aur packages on arch, we could comments package and mark outofdate
<eoli3n_>
how do i do this with a nix package ?
dozn_ is now known as dozn
<PyroLagus>
eoli3n_: open an issue on the githubs
<eoli3n_>
PyroLagus: but how will the specific maintener be notified
<PyroLagus>
there's a bot that notifies everyone who's in the maintainers attribute
<eoli3n_>
oh ok cool :)
<PyroLagus>
if the issue is formatted correctly, i think
takeda has joined #nixos
<eoli3n_>
Should i open a bug report or a packaging requests
<eoli3n_>
about the bot, you "know" or you "think" ?
<eoli3n_>
:)
<PyroLagus>
hm. maybe the bot doesn't notify for issues. i'm not sure now…
<eoli3n_>
maybe you're right
<eoli3n_>
in md file there's a specific mainter information field
<eoli3n_>
maintainer
<eoli3n_>
lets try
<PyroLagus>
that said, it DOES notify for PRs, sooooo you could just make a pr where you just change the version number and hash. there's a good chance it would work even
<eoli3n_>
hm, i started nixos 2 days ago, i still don't know where i should modify that
<eoli3n_>
and i see 1707 pull requests on nixpkgs, wow, that's huge
takeda has quit [Ping timeout: 258 seconds]
LLED2_2 has joined #nixos
o1lo01ol1o has joined #nixos
ouzu[m] is now known as ouzu
dvim has joined #nixos
ouzu has quit [Quit: authenticating]
LLED2_ has quit [Ping timeout: 268 seconds]
<DigitalKiwi>
eoli3n_: what is out of date ?
<DigitalKiwi>
if you want i'll try and update it and do a PR
<etu>
Izorkin: Thanks. I put it on my list of nixos things to look at for tomorrow evening or the weekend. That list is new and short except of the PineBook Pro that I will pick up today. But I should get to your PRs :)
<eoli3n_>
i want to verify if font-awesome is the otf version
zupo has joined #nixos
felixfoertsch has quit [Quit: ZNC 1.7.4 - https://znc.in]
tomberek has joined #nixos
<Shoubit>
eoli3n_, might not apply to what you're doing but if you're looking for something specific you can use nix-locate (from the nix-index package) e.g. nix-locate '.otf' | grep -i fontawesome
chagra has joined #nixos
felixfoertsch has joined #nixos
<eoli3n_>
ok thx
<FRidh>
Building pkgsStatic.openblas results in /nix/store/m8ih8gw7qnd2c7p0qgv3cn4b6abkbanb-binutils-2.31.1/bin/ld: cannot find -lc
<clever>
FRidh: it could be forcing a dynamic build, or ignoring the static config flags
miguelnegrao has joined #nixos
<{^_^}>
[nixpkgs] @misuzu opened pull request #76817 → nixosTests.3proxy: port test to python → https://git.io/Jexur
<FRidh>
clever: any suggestion on how to debug this, aside from going through the makefile?
<FRidh>
mmahut: maybe that package at master is different than that package at the nixpkgs version
<{^_^}>
[nixpkgs] @andir opened pull request #76820 → buildRustCrate: add `buildTests` flag to tell rustc to build tests instead of binaries → https://git.io/Jexup
<dirkx>
Where can one see the output of a `build' request to @GrahamcOfBorg -- or see the result of these builds ? Or is that something private ?
<mmahut>
FRidh: it is not, that's the odd thing, I'm trying to debug this for hours now :)
<coco>
for the option "services.xserver.displayManager.lightdm.background", how can i provide a custom background? assigning a path doesn't seem to work
<alex``>
leotaku: hey o/
<alex``>
how to make xdg-open available?
cosimone has quit [Quit: Terminated!]
<alex``>
I just switched to Wayland, removed my X11 settings, not sure which setting was responsible of `xdg-open` being available in PATH
smollcato has quit [Remote host closed the connection]
noonien has joined #nixos
<smollcato>
[telegram/nixos] <SchmollCatto> test
smollcato has joined #nixos
<gchristensen>
that message did come through, but it came through quite verbosely
<gchristensen>
<smollcato> [telegram/nixos] <SchmollCatto> test
<noonien>
is it possible to add more packages to an existing derivation? for example, i have polybar that runs some scripts, i want to have `jq` for those scripts, without installing it system-wide, nor user-wide
<chloekek>
noonien: you can wrap a group of derivations in a new one with buildenv.
<noonien>
yeah, i think that's what i'm looking for!
felixfoertsch has quit [Quit: ZNC 1.7.4 - https://znc.in]
globin_ is now known as globin
globin has joined #nixos
globin has quit [Changing host]
ebzzry has quit [Ping timeout: 268 seconds]
cfricke has joined #nixos
<chloekek>
Of course the PATH entry should be ${jq}/bin, not ${jq}/bin/jq.
felixfoertsch has joined #nixos
ebzzry has joined #nixos
<chloekek>
Usually I even go as far as using --set PATH instead of --prefix PATH, to ensure it’s truly self-contained. But that may not be applicable for your use case (or it may be).
<noonien>
i think it will be, this is just a simple example
<noonien>
for polybar atleast, i'll have to install some language servers for nvim later, that'll probably get a bit tricker, but probably not by much
<chloekek>
makeWrapper adds a layer of indirection, and is therefore able to solve all problems in computer science.
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<noonien>
hopefully post-makeWrapper cryptography takes off until more people realise this
<{^_^}>
[nixpkgs] @mdlayher opened pull request #76825 → corerad: init at v0.1.1 → https://git.io/JexgA
<ryantm>
The Final Comment Period for RFC 0046 has started with disposition to merge and, barring any blocking issues, will be merged after 2020-01-12. Your opinions, comments, and approvals are welcome! https://github.com/NixOS/rfcs/pull/46
<acowley>
I have an issue with an update to nixos-rocm that I'm trying to package: OpenCL-using programs link to libOpenCL.so that we provide, but in the latest version, OpenCL-using programs fail when starting up. Using strace, I found that something is trying to open an additional .so that we build as part of a separate package. If I set LD_LIBRARY_PATH to include /nix/store/other-package/lib, then OpenCL programs work. But I haven't been able
<acowley>
to track down *what* is trying to open that .so so I can give it an absolute path. Can anyone suggest a way of doing that?
cosimone has quit [Quit: Terminated!]
<{^_^}>
[nixpkgs] @mmahut merged pull request #76814 → fail2ban: fix work with python3 → https://git.io/Jex03
<acowley>
An alternative would be a way of doing the same thing as the LD_LIBRARY_PATH "fix" in such a way that it happens automatically. I think AMD's testing doesn't catch this because they just dump everything in /usr/local/lib or some other path that their testing OS always searches when trying to open a dynamic lib.
<chloekek>
What usually happens is that the dynamic loader opens the .so file, and the absolute path is in the ELF file.
<chloekek>
AFAIK LD_LIBRARY_PATH causes the dynamic loader to take the basename of that hardcoded path and search it in LD_LIBRARY_PATH instead.
<chloekek>
You can change the dynamic loader with patchelf (I believe it’s the interpreter setting) but I don’t know the details.
<chloekek>
I think you can also change the absolute path with patchelf.
<chloekek>
If they use dlopen then it’s another story, of course.
<acowley>
chloekek: Which file are you suggesting that I patchelf to set the absolute path?
ng0_ has joined #nixos
ng0_ has joined #nixos
ng0_ has quit [Changing host]
<chloekek>
The executable.
<ptrcmd>
lewo`: Thanks!
domogled has quit [Quit: domogled]
tsrt^ has quit []
<chloekek>
The programs that use OpenCL, that is.
<chloekek>
It may be easier to just set LD_LIBRARY_PATH with makeWrapper or wrapProgram though, if patchelf doesn’t work out and you don’t mind writing a derivation.
<acowley>
chloekek: It doesn't mention this additional lib (comgr) in ldd. I think they are using dlopen, because they do this thing where they check for the availability of optional proprietary components.
<acowley>
But I simply haven't been able to find it!
<chloekek>
acowley: what does grep say if you grep the executable for the absolute path?
<chloekek>
acowley: and what if you grep it for the basename?
<acowley>
It's a bit non-trivial because when they do use dlopen, they tend to construct the path name for cross-platform compat, but there aren't that many uses of dlopen, so I've reviewed all the ones I could find.
<the_pumpkin_man[>
Hello, I'm actually writing a systemd service. It is running under a certain system user. The problem is I want to execute the preStart script as root, not as the user. How can I do this wih NixOS ?
dansho has quit [Ping timeout: 265 seconds]
ng0 has quit [Ping timeout: 268 seconds]
<acowley>
chloekek: The missing lib isn't mentioned in the executable. The wrapper approach is a viable fallback, but not at all ideal as it would mean that OpenCL-using programs need wrapping.
ng0_ is now known as ng0
<chloekek>
Well you need to either wrap them or patch them.
<the_pumpkin_man[>
I have seen that but the problem is Im using the NixOS's implementation
<the_pumpkin_man[>
Im defining it in the configuration.nix file
<the_pumpkin_man[>
with systemd.services.<name>.preStart
<chloekek>
acowley: I’m not sure how Nixpkgs generally deals with propietary software. I think it’s rather ad-hoc. (For instance, Steam is run in a chroot because only proprietary Debian-compatible packages are available.)
<notgne2>
the_pumpkin_man[: if you don't find another solution, you could make it be a root service and drop to a user using su
<the_pumpkin_man[>
notgne2: Or just make another service running as root, which will be loaded before my main service
<notgne2>
chloekek: the Steam chroot afaik is because binaries for NixOS must be patched to run, and patching every steam game is an unachievable goal
<eon`>
you can set systemd.services.<name>.serviceConfig.ExecStartPre directly so that you can add '+' before the command
<the_pumpkin_man[>
But I want to find a convenient way to do it
<acowley>
chloekek: We've done okay navigating this in the past. I make the relevant proprietary lib an optional download where I unpack a .deb they provide. This new lib isn't actually proprietary, so if they've wrapped it in their very-dynamic plugin finder, then that seems to havbe been a mistake.
<the_pumpkin_man[>
Oh eon` good idea will try
<{^_^}>
[nixpkgs] @fpletz merged pull request #76065 → qemu: 4.1.0 -> 4.2.0, including patch for minor CVE-2019-15890 → https://git.io/Je5H6
<Palpares>
hi. i am stuck with the nixos proxy config. my config file had a configured proxy that worked. now i changed my netword and wish to remove the proxy. I edited the configuration.nix and commented the proxy. howver when I run rebuild- the system still consider the proxy and I get "could not resolve the proxy" timeout. How to remove a proxy once configured ?
mexisme has joined #nixos
cantstanya has quit [Ping timeout: 240 seconds]
iwvt[m] has left #nixos ["User left"]
rootatarch has joined #nixos
elux has joined #nixos
szicari has joined #nixos
<elux>
hi all
gm38 has joined #nixos
cantstanya has joined #nixos
<rootatarch>
bye
<elux>
im installing a new nixos system with full disk encryption. do i need to use grub? ive been reading that grub doesn't support luks2 which i'd like to use to stay current
<rootatarch>
grub supports it
mexisme has quit [Ping timeout: 260 seconds]
<rootatarch>
check out arch wiki
<elux>
ok thanks, perhaps that is outdated. yea good idea, i'll check arch wiki - its usually pretty detailed
<gm38>
Has anyone managed to get their RX 5700XT working on NixOS? I'm struggling, the system boots but it just displays a blank screen
<rootatarch>
intel and nvidia guy here
<elux>
gm38: have you tried the latest linux kernals? 5.x ? i hear that is when 5700 support came about
<rootatarch>
grub does not supports luk2 dude to somebug
<{^_^}>
#68809 (by jb55, 15 weeks ago, closed): Get navi10 (RX 5700 XT) amdgpu working
<nilsirl[m]>
I'm testing some stuff and I would like to have write access to /nix/store
<nilsirl[m]>
Would that be possible?
b1000101 has joined #nixos
<exarkun>
I am suddenly and mysteriously unable to open luks devices as non-root
<exarkun>
to make it more interesting, I will also throw in the traditional "nothing on the system has changed since it worked".
<exarkun>
who wants to wow me with a procedure to magically fix it
<rootatarch>
try rebooting
<rootatarch>
:D
<exarkun>
rebooting is so old hat is has moved from the realm of the magical to that of the mundane
<exarkun>
but I guess I could nevertheless give it a try
<chloekek>
You probably need to be in some group that you need to be in.
<chloekek>
That’s what “suddenly” stopped audio from working for me.
<exarkun>
I'm in all the groups today I was in last week
<gustavderdrache>
don't forget that rebooting replenishes your CPU's supply of magic bits
takeda has joined #nixos
<exarkun>
I should apply the kernel patch for hot magic-bit-replenishment someday
<gustavderdrache>
one of my company's sysadmins lives near an enchanted forest and they just bring some moss back and sprinkle it on our critical systems to achieve the same effect
<veleiro>
anyone working with nixos on the pinebook pro?
<gchristensen>
I think yeah, #nixos-aarch64
<veleiro>
cool
jgeerds has quit [Ping timeout: 240 seconds]
xelxebar has quit [Remote host closed the connection]
<MmeQuignon>
Hi ! Is there someone familiar with "next" browser ? I'm trying to install it but it complains about not finding "/nix/store/qm5yk280d558im7x8l59gr82ipnq6llw-next-gtk-webkit-1.3.4/bin/next-gtk-webkit"
<MmeQuignon>
should I install something else to make it work ?
tsrt^ has joined #nixos
<lewo`>
MmeQuignon: no, it is supposed to work out of the box.
<lewo`>
MmeQuignon: maybe you can try with another nixpkgs channel
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<MmeQuignon>
I'm on unstable and not that familiar with pinning
<dminuoso_>
MmeQuignon: How did you install it? Can you please pastie the full output (including your command)?
<the_pumpkin_man[>
eon`: misuzu Thank you both it worked !
<the_pumpkin_man[>
Can I also do this for example : ExecStartPre = "test ! -d /dir && +mkdir /dir"
<the_pumpkin_man[>
?
<MmeQuignon>
next is a browser which seems nice, I just want to try it
<MmeQuignon>
I installed it in users.users.matthieu.packages
<Shoubit>
MmeQuignon, you don't necessarily need to pin (although it's probably the least stateful), you can do it on a per-channel basis if that's enough with: nix-env -f channel:nixos-19.09 -i next
<lewo`>
MmeQuignon: I reproduce your issue.
erictapen has joined #nixos
<MmeQuignon>
Shoubit: It's not in the stable channel (I think)
<zeta_0>
i want to have the path automatically shortened, like i have in elisp in emacs:
<zeta_0>
[zeta@nixos:/nix-haskell-recipes]$ pwd
jmeredith has quit [Quit: Connection closed for inactivity]
<immae>
Hello there! I lost ipv6 connectivity (I have an ip6 adress defined correctly via `interfaces."ens3".ipv6.addresses`, but I’m unable to ping remote hosts) when upgrading from nixos-19.03.173017.85f820d6e41 to nixos-19.03.173677.daf861a810d (with linuxPackages_latest), do you know of any issues/changes recently? It did it consistently on two servers.
chloekek has joined #nixos
<rnhmjoj>
immae: is your server a linode vps by any chance?
<immae>
rnhmjoj: no a hetzner server (VM)
<immae>
Both servers are in the same datacenter, so it could be a general issue, but since it happened *right after* restarting I think I’m in fault :p
<rnhmjoj>
immae: i asked because linode blocks ipv6 privacy extensions wich are enabled and used by default in 19.09
<immae>
ah
<immae>
I’m still in 19.03
<rnhmjoj>
however that changed in 19.09, not in 19.03
<immae>
If it helps (I lack knowledge in ip6), a `ping -6 fe80::%ens3` "works"
xkapastel has quit [Quit: Connection closed for inactivity]
erasmas has joined #nixos
<zeta_0>
rycee: i am not sure if this is a home-manager problem, that is why i am asking you for help?
averell has quit [Quit: .]
<immae>
rnhmjoj: Just fount it, it turns out I simplly missed a `ip -6 route add default via fe80::1 dev ens3`... I don’t understand how it used to work before...
averell has joined #nixos
rogue_koder has joined #nixos
domogled has joined #nixos
<rnhmjoj>
uhm, that's strange... i gave a quick look at the diff between the versions you mentioned but i didn't find anything related to ipv6 or even networking
<{^_^}>
[nixpkgs] @ericnorris opened pull request #76830 → Disable NS in emacs-nox → https://git.io/JexrW
<immae>
rnhmjoj: well, I’ll add it explicitly to my config :) I found networking.defaultGateway6, is that the right way to add my route?
sigmundv_ has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @mmahut merged pull request #76817 → nixosTests.3proxy: port test to python → https://git.io/Jexur
<rnhmjoj>
immae: if you configured the network manually with that command the changes will probably be lost when switching to a new nixos configuration. you should use the nixos config to make that persistent
<immae>
rnhmjoj: I did it on one host to "check", and on the other host I did it via nixos config
<{^_^}>
[nixpkgs] @mbrgm opened pull request #76831 → <!-- Nixpkgs has a lot of new incoming Pull Requests, but not enough people to review this constant stream. Even if you … → https://git.io/Jexom
<nix-newbie>
Hi all. Trying out nix for a server deployment. Veeeery green, but would like to learn. Specifically, adding multiple users in configuration
<nix-newbie>
I know I can do
<nix-newbie>
users.users.X = { ... };
<nix-newbie>
But say I have 20 users to create with the same configuration
<nix-newbie>
That would seem to do be more than nice enough.
<nix-newbie>
It worked, and it is beautiful
<nix-newbie>
Thank you
fresheyeball has joined #nixos
<infinisil>
:D
<infinisil>
I personally have a thing where i can more or less do `users.mainUserConfig = { <the user config> };` and `users.mainUsers = [ "user1" "user2" ]`, and it applies the config to all main users
<infinisil>
And I can add main users or add to the config from anywhere and it still applies to the them all
zeta_0 has joined #nixos
<nix-newbie>
I'm trying to get a system off the ground to demonstrate the superiority of this over our Ubuntu + Ansible setup which is giving nigtmares.
<infinisil>
Hehe nice
<nix-newbie>
I'm sure I will have an itch to get fancy down the line
<zeta_0>
i am able to rebuild switch this without any errors, but it still does not shorten the path of the bash shell?
<thblt>
Hello! Is there a process to refresh the list of derivations automatically generated from MELPA/ELPA Emacs packages? Like an auto rebuild every n months or something?
growpotkin has joined #nixos
<rnhmjoj>
zeta_0: the slashes are being gobbled up by the nix interpreter
<rnhmjoj>
zeta_0: you should use the '' quoted string to avoid it
<turq>
trying to build a haskell project using cabal new-build within a nix shell and I see that zlib-0.6.2.1 requires a build, but when it goes to build, it can't find the header. there's a section in the nixos manual about stack complaining about missing system libraries, but I didn't have any luck
<clever>
zeta_0: try opening another terminal window
dm9 has quit [Quit: WeeChat 2.7]
kuznero has joined #nixos
<turq>
/nix/store only has zlib-0.6.2.1.drv and zlib-0.6.2.1tar.gz.drv
<kuznero>
Hi All!
<buckley310>
zeta_0 sessionVariables are set by PAM, so that would be before your shell is started. when your shell is started, it will source /etc/bashrc, which overrides PS1. try setting PS1 in programs.bash.interactiveShellInit instead
<Ericson2314>
p01ar: oh neat! what sort of Windows problems?
<buckley310>
zeta_0, actually, just found a dedicated setting for this: programs.bash.promptInit
<kuznero>
I am trying to `nix-shell` having [this](https://pastebin.com/au8FqzCQ) in my shell.nix file. It fails to `cargo install diesel_cli` when ran from `shellHook`, but when I run it myself once inside the shell, it runs fine. Why is that?
<turq>
in short: i can nix-build just fine, but i can't cabal new-build due to the lack of a library
<kuznero>
It seems that it is somehow related to the fact that I have rustup installed the ordinary way...
<kuznero>
:(
selfsymmetric-pa has quit [Remote host closed the connection]
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
<ouzu>
Does it even make sense to use btrfs with nixos? I currently use it for snapshots and on nixos I could just put the configs into a git repo right?
<lassulus>
for /nix/store it doesn't really make sense, for other stuff it depends on your usecase
<gchristensen>
lassulus++
<{^_^}>
lassulus's karma got increased to 5
<buckley310>
btrfs keeps data checksums, which is nice. it also allows you to use subvolumes instead of partitions, so you dont have to explicitly allocate space for, for example, a home partition.
<buckley310>
that being said, nixos has excellent zfs support as well.
<benley>
I use ZFS for /nix/store and that has been working great
<benley>
I imagine btrfs would be useful for similar reasons
jgeerds has joined #nixos
<benley>
(checksums for data integrity, primarily)
<benley>
and compression, if btrfs supports that
<ouzu>
ok that makes sense
<ouzu>
yeah, it supports at least deduplication
<benley>
oh yeah it does. My /nix/store volume is compressed at 2.1 to 1 with lz4
<benley>
I don't use dedup because it requires a ridiculous amount of ram and nix can do that by itself anyway
<benley>
zfs dedup, that is
<ouzu>
ok thanks!
<nahamu>
+1 for using ZFS. NixOS support for it is great and ZFS is great.
<ouzu>
zsf even has an entry on the wiki :)
<benley>
nixos zfs support continues to delight me :-)
<gchristensen>
ouzu, benley: dedup is a bit dangerous, #zfsonlinux exclusievly recommends *against* it for anyone who asks about it
<gchristensen>
("if you have to ask, you don't want it" is sort of their approach)
<benley>
yeah - I wouldn't recommend it for most cases even if it were guaranteed safe. It's extremely resource intensive.
<gchristensen>
and if your dedup table stops fitting in RAM your optios are basically buy more ram, or restore from a backup from before it got too big
<clever>
gchristensen: ive never run into issues where it doesnt fit, but ive not done full dedup on multi-tb datasets yet
<benley>
clever: is that zfs or btrfs?
<benley>
(I know btrfs has some dedup stuff but I know almost nothing about it)
<clever>
benley: zfs
<buckley310>
dedup on btrfs is interesting. its safe to try and reversible if its no good.
<buckley310>
they basically punt to userspace, and whatever process in userspace scans the drive and tells the kernel which references to share
<gchristensen>
oh cool
<gchristensen>
! you could implement nix's hardlink stuff there!
<clever>
buckley310: ah, sounds like xfs_fsr's defrag tools
<clever>
basically, its a userland tool that will copy an entire file using standard api's, then an ioctl that will atomicly swap the backing blocks of 2 files, and check for race conditions (source got modified)
<clever>
then it just relies on the fs to make better choices for the copy of the file
<benley>
that's awesome
<buckley310>
ah, thats cool. i didnt know xfs could do that type of dedup
<clever>
buckley310: thats less of a dedup, and more of a defrag
cosimone has quit [Quit: Quit.]
<clever>
it uses the filefrag api (and binary by the same name) to just get extend lists, so it can count the fragments
<buckley310>
it sounds like you can take two files that share extents and tell xfs to share them retroactively, just a cursory look though...
<clever>
and if over a set limit, it just duplicates the entire file, and swaps the backing blocks if fragmentation improved
<lovesegfault>
infinisil: Niteo link is borked on the README
<infinisil>
Whoops
<buckley310>
zeta_0, you want a very basic PS1 variable all the time, right?
<infinisil>
E.g. I could write a multi-host module that allows me to select a server and a client, and it does all the configuration for the server and client machine
<zeta_0>
buckley310: but i think it is missing the `/` that is supposed to be placed before the currently working directory path:
<zeta_0>
[zeta@nixos:`/`nix-haskell-recipes]$ pwd
<zeta_0>
buckley310: ^
<gchristensen>
lovesegfault: (it hasn't run in Some Time Now but ...) r13y.com is basically a loop over a nixos system and running nix-build with --check :P
<lovesegfault>
gchristensen: Oh wow
<lovesegfault>
that's pretty cool
<gchristensen>
I have hopes to get it going again tomorrow
<zeta_0>
buckley310: now i just need to figure out how and where to place the `/` in the code and i'll be good to go
<lovesegfault>
gchristensen: Why did it stop?
<lovesegfault>
gchristensen: Eh, is GCC (or clang) deterministic?
<{^_^}>
[nixpkgs] @nlewo opened pull request #76838 → next: set next-gtk-webkit as a propagated build input → https://git.io/Jexib
<gchristensen>
no it isn't dead, it just ... can't do this anymore, haha.
elux has joined #nixos
<lovesegfault>
:D
<NoctisLabs>
Steam recently started crashing, I think it has something to do with the hardware survey window that's popping up. Here's the log: https://paste.debian.net/hidden/a2c6b966/
<gchristensen>
it is very close to death, but it still streams plex well enough that I haven't pitched it :)
<lovesegfault>
gchristensen: I'd be curious to see whether there are any changes between x86 and aarch64 in how reproducible it is
<gchristensen>
a good question
mexisme has joined #nixos
<lovesegfault>
is the -ck kernel packaged for NixOS?
<{^_^}>
[nixpkgs] @aszlig merged pull request #76697 → nginx: Clear Last-Modified if ETag is from store → https://git.io/JeAmx
<lovesegfault>
it does include someone stealing my laptop and getting hold of my documents
<gchristensen>
lovesegfault: all of my encryption is maleable with enough wrenches
<elux>
my threat model would be to extent someone gets physical access to computer while its turned off
<lovesegfault>
elux: So you want to be protected against evil-maid type attacks?
<lovesegfault>
That's going to be... tough
<lovesegfault>
gchristensen: You use BIOS/GRUB/LUKS1, right?
<elux>
do you guys know how the nvme hardware encryption works? i hear Samgsung drives are encrypted by default and if you set a bios password it will use that as the key? is that true...?
<lovesegfault>
elux: It's called OPAL
<NoctisLabs>
I don't use encryption on devices in my home, I only use encryption on devices that leave my home.
<gchristensen>
elux: that is true, but it afaik it is for chips inside and if you plug it in to anything it happily decrypts
<clever>
elux: some drives actually use the pw you set to encrypt, others are lazy and just use the password as a boolean to allow/deny
<clever>
elux: ive heard that some drives actually do things a bit like luks, its ALWAYS encrypted, and setting a pw just encrypts the master key
<elux>
gchristensen: happily decrypts so long as someone knows the passphrase?
<gchristensen>
my understanding is it didn't take a user-defined pw
<elux>
yea i was told the same thing clever. but im wondering if its true
<gchristensen>
but that is something else, sorry
<clever>
elux: there are also some drives, that dont directly involve the bios, on first boot, it shows a dummy plaintext partition, you must then install software to that, which will run the unlock tool, and chainload the real os
<gchristensen>
o.o
<clever>
so it will claim to be a 256mb drive, but if you run a util, it magically grows, and the contents are swapped out
<elux>
yea this is why im installing luks to be safe, but, turning on that feature at bios level can presumably give some extra protection without any overhead to the performance since its just changing the master key
<gchristensen>
I wouldn't trust hard disk's internal encryption support :)
<elux>
roger that :)
<gchristensen>
there was a whole nightmare article about this
<elux>
gchristensen: do you encrypt your /boot as well? or just everyting else?
<gchristensen>
I don't encrypt /boot
<gchristensen>
one of my laptops uses secureboot though
zeta_0 has quit [Remote host closed the connection]
zeta_0 has joined #nixos
<{^_^}>
[nixpkgs] @aszlig pushed to release-19.09 « nginx: Clear Last-Modified if ETag is from store »: https://git.io/JexPi
justanotheruser has joined #nixos
<zeta_0>
i am finally am to shorten bash's path but when go into a nix-shell, the path gets long again, why doesn't the shell in nix-shell enherit the bash configuration that i set in configuration.nix(programs.bash.promptInit)?
<zeta_0>
as you can see, having a long path can get very annoying
<elux>
nice. efi, luks2 and systemd-boot -- just did an install and its working like a charm :) so much better than awful grub
cab404 has quit [Ping timeout: 260 seconds]
<multun>
FRidh: hello :) in this pull request, you said you do not intend to submit any newer sphinx until after the 19.09 release, do you think it's relevant to do it now? https://github.com/NixOS/nixpkgs/pull/67730
<{^_^}>
#75773 (by worldofpeace, 2 weeks ago, merged): gegl_0_4: add patch for test timeouts
<lovesegfault>
I'm sick of gegl failing and ruining the release
<lovesegfault>
can I just patch out that damn test
<lovesegfault>
Today we won't get a release because of gegl failing again :(
justanotheruser has quit [Ping timeout: 260 seconds]
nexgen2 has quit [Quit: Leaving]
klimi has joined #nixos
<klimi>
Hello, how can i change default app with xdg-mime? No .desktop files are found, in pcmanfm i can chose to open the pdf in evince but if i chose the "Open with" option there are none apps listed so i can not change it. Any help appriciated
<klimi>
oh so i figured it out... the correct name is org.gnome.Evince.desktop...
<klimi>
infinisil: Thanks
<klimi>
And can i somehow set this in my configuration.nix?
eraserhd2 has joined #nixos
<immae>
Hello there, until today I was convinced that "${./foo}" was absolutely equivalent to ./foo, but just now I used that and in the derivation ./foo became the path to my local directory instead of a path to a nix store containing the content of the directory, is that expected? (FWIW I used it as `WorkingDirectory = ./foo` in a systemd service definition. `WorkingDirectory = "${./foo}"` worked fine but I’m
<immae>
puzzled)
<infinisil>
klimi: Not really, this is user-level stuff. You could still do it with some script that runs on startup, but that's kind of ugly
eraserhd has quit [Ping timeout: 260 seconds]
thomashoneyman has joined #nixos
<infinisil>
immae: "${./foo}" imports the path into the store and returns the store path
justanotheruser has joined #nixos
<thomashoneyman>
i'm attempting to back up some files to a USB drive, but i've only now realized i have no idea how to do this in NixOS
o1lo01ol1o has joined #nixos
<infinisil>
immae: Wheras ./foo is the absolute path directly
<immae>
hmm
<thomashoneyman>
anyone have a link to docs on this or know how?
<thomashoneyman>
just trying to mount the usb, copy the files, eject
<infinisil>
immae: If the module that reads these values does "${value}" then even when `value = ./foo` it still gets imported into the store
<infinisil>
immae: However if that's not the case, then it won't get imported
<immae>
ahh
<infinisil>
So depending on what the module does, ./foo can be imported or not
<thomashoneyman>
unfortunately my google-fu isn't great because of how many results are about using a usb to install nixos
<infinisil>
And systemd's module does the latter
<immae>
infinisil: ok, so it’s my first occurence of a module not doing that :D
<immae>
Thanks
<infinisil>
:)
<klimi>
infinisil: so home manager with xdg.mimeApps... Thanks anyway
<infinisil>
thomashoneyman: If you don't use an automatic mounter, you can do `sudo mkdir /mnt && sudo mount /dev/sdX1 /mnt` then just `cp -r some-directory /mnt` then `sudo umount /mnt`
<infinisil>
klimi: Ah nice, didn't know about that
<thomashoneyman>
infinisil: i think i'd rather use an automatic mounter, just never heard of them
zeta_0 has joined #nixos
<thomashoneyman>
thanks for the note
klimi has quit [Quit: WeeChat 2.6]
<infinisil>
Well I'm not using one, so no idea, maybe some DE's do that
<thomashoneyman>
have one you prefer?
<thomashoneyman>
oh, i see
<thomashoneyman>
i just have i3 going, so there's no "desktop" per se
<jackdk>
if you're using home-manager then turning on the udiskie service is easy, and I think you need services.udisks2.enable = true; in configuration.nix
<thomashoneyman>
infinisil: hm -- that's giving me "wrong fs type, bad option, bad superblock on /dev/sdc, or other error"
<infinisil>
thomashoneyman: Well you need to give the correct partition
<infinisil>
/dev/sdc<whatever number>
<zeta_0>
how would i get nix-shell to inherit from my bash setting in configuration.nix? is there a simple way or do i have to setup something like `direnv` ?'
<thomashoneyman>
sorry, i haven't done this before (quite new to linux in general)
<thomashoneyman>
i'm assuming i'd see that from fdisk
<thomashoneyman>
checking...
<thomashoneyman>
mm i see a "Disk /dev/sdc" and also a "Device /dev/sdc1"
<thomashoneyman>
that's better: "unknown filesystem type 'exfat'"
<thomashoneyman>
(using /dev/sdc1)
<thomashoneyman>
ok i'm looking at mount docs now
<thomashoneyman>
should be ok, thanks for the tip!
<infinisil>
Or `lsblk` instead of `fdisk`
<infinisil>
Much better output for lsblk
<thomashoneyman>
infinisil: yes, that's quite a bit better
<thomashoneyman>
just have to check how to mount exfat
<Ashy>
hmm, `python` is still version 2.7.17 for me but pkgs.python or pkgs.python27 aren't referenced in my configuration.nix or in `nix-env -q`, how do i figure out where this version of python is coming from?
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Ashy>
not using home-manager and i don't currently have my configuration repo public, what configuration.nix options would update the PATH?
<infinisil>
Hm no idea
<infinisil>
Ashy: Okay how about this: Run `sudo htop`, go to the process of your terminal, press "e" to see the environment variables of the process, check whether PATH contains those /nix/store paths
<logzet>
Or my original question: Is it possible to have multiple packages, that depend on each other, in one .nix file? And if it is: is there an example to look at in nixpkgs?
selfsymmetric-pa has joined #nixos
<ajs124>
logzet: they are in all-packages.nix. I'd send you a link to github, but it seems to cut off the file
<logzet>
I didn't find them in all-packages.nix because github cuts the file ^^
<logzet>
thanks :)
selfsymmetric-pa has quit [Remote host closed the connection]
selfsymmetric-pa has joined #nixos
<Ashy>
/etc/profile and /etc/bashrc both look ok, or at least aren't explicitly setting $PATH from what i can see
<Ashy>
infinisil: i didnt know about 'e' in htop that's nice
<selfsymmetric-pa>
Oh that IS nice. infinisil++
<{^_^}>
infinisil's karma got increased to 174
<Ashy>
so the /run/current-system/sw/bin/bash PATH envvar already has the nix store paths in it and so does the terminal above that in the process tree
<clever>
Ashy: you can also peek at /proc/PID/environ to see env vars, its null seperated
<Ashy>
the terminal is a wrapper though: /nix/store/ry7pmp6qc5r3gl7dk3ha3swyankv2hx2-python-2.7.17/bin/python2.7 /nix/store/3r5mwr1rk8aqip0km4nw6cczkm335526-terminator-1.91/bin/..terminator-wrapped-wrapped
selfsymmetric-pa has quit [Remote host closed the connection]
<Ashy>
htop with t and e is pretty interactive to scan around with though
<Ashy>
i think i'm on to something possibly, fired up an xterm and ran htop and PATH envvars in that tree have far less direct nix store references compared to the terminator processes
<infinisil>
Hm yeah so maybe it's terminator doing something
<infinisil>
Also very suspicious because it's using python2
<zeta_0>
clever: quick question, how do i configure nix-shell? would i need to inherit from the bash settings in configuration.nix using direnv, or is there any easier/simpler way
vidbina has joined #nixos
<zeta_0>
clever: everything in nix-shell works fine, the only problem that i have is the long path, which is annoying
<{^_^}>
[nix] @edef1c opened pull request #3297 → passAsFile: hash the attribute name instead of numbering sequentially → https://git.io/JexM5
<tracer41matrix[m>
zeta_0: PM?
<clever>
zeta_0: by default, nix-shell will add to most env vars
<gchristensen>
nice, edef
<{^_^}>
[nix] @edolstra merged pull request #3297 → passAsFile: hash the attribute name instead of numbering sequentially → https://git.io/JexM5
o1lo01ol1o has quit [Remote host closed the connection]
<lovesegfault>
that was quick
<tracer41matrix[m>
clever: sounds good!
<niksnut>
:-)
<edef>
niksnut: <3
<tracer41matrix[m>
but simple ... :[
<puck>
that was FAST
cosimone has quit [Quit: Quit.]
shiver has joined #nixos
o1lo01ol1o has joined #nixos
<lovesegfault>
gchristensen: fixed
<lovesegfault>
gchristensen: With that fix merged can you manually spin a new release or do we have to wait another day?
fresheyeball has quit [Read error: No route to host]
<gchristensen>
lovesegfault: if there is an upstream bug report, it'd be good to put a link to it in the comment in the expression (if not, it might be good to make a bug report upstream). it triggers every 4h, but sure we could trigger an eval once it merges
<lovesegfault>
Alright, let me look upstream
<gchristensen>
thank you!
<zeta_0>
clever: i would like nix-shell to have this configuration, it basically shortens the path to the current working directory as well as a few other things? is there some setting where i could do this or would it have to be inherited from configuration.nix using something like direnv?
<clever>
zeta_0: i would just put that into the shellHook for your shell.nix
<gchristensen>
awesome!
<gchristensen>
lovesegfault: then yeah, once it passes ofborg we'll merge and trigger an eval :)
<gchristensen>
lovesegfault: thank you for hunting this down and patching
<lovesegfault>
gchristensen: No worries :D
<kraem>
i can't use wifi after suspend, only if i restart iwd.service (on both LTS and mainline kernel). where should i start searching?
<lovesegfault>
I'm building on my server too
<zeta_0>
clever: does the name of the nix file matter, i am currently using a default.nix file?`
<clever>
zeta_0: nix-shell will load shell.nix by default, and fall back to default.nix when shell isnt found
<gchristensen>
lovesegfault: is there a different gegl I should try building? ofborg seemed to be able to substitute it
<lovesegfault>
gchristensen: it's gegl_0_4
<zeta_0>
clever: oh i see, i was not sure if you meant that there was a global shell.nix file somewhere, so in whatever project root i am in, i need either a default.nix or shell.nix file in the root directory to place that piece of bash code, is this what you mean?
<clever>
zeta_0: yeah
<tracer41matrix[m>
hi, im tracer
o1lo01ol1o has quit [Remote host closed the connection]
<gchristensen>
hey, tracer41matrix[m :)
<judson_>
Is there an updated version of signal-desktop?
<Valodim[m]>
I'm wondering about Python packaging, perhaps someone can enlighten me: am I reading it right that all python library dependencies must be packaged and included as nix packages, but applications are then built "statically" by including all the python library dependencies?
<clever>
zeta_0: ive not looked heavily into direnv yet
lukash_ is now known as lukash_|away
<zeta_0>
clever: ok, i will try trial and error with `direnv` to see if i can get it to work
<zeta_0>
clever: what is a shellHook, i don't think i have ever used that before?
<lunik1>
I am trying to switch from NetworkManger to connman, and I have added networking.connman.enable = true to my configration. But now wpa_supplicant fails to start because there is no /etc/wpa_supplicant.conf. What am I missing?
<clever>
zeta_0: if you run nix-shell on a derivation that has a shellHook, then nix-shell will run whatever bash code is in the shellHook, before giving you a shell
alexherbo2 has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<zeta_0>
clever: one more question, do the programs.bash options used in configuration.nix work in a shellHook as well? or will the code have to be different?
<clever>
alexherbo2: nixos will re-create it on boot, but your nix-env profile and nix-channel profile will be gone
zaeph has quit [Ping timeout: 252 seconds]
<clever>
zeta_0: the env vars nixos had set, will mostly still be set in shellHook, but nix-shell may have overwritten some and prepended ot others
<lovesegfault>
gchristensen: Woohoo! trigger ze hydra und ship eeet
<gchristensen>
lovesegfault: an evaluation is pending :)
<tracer41matrix[m>
i may be friendly but i wont make your or alexherbo2s homework; gchristensen
<lovesegfault>
tracer41matrix[m: What homework? What are you talking about?
* lovesegfault
high fives gchristensen
vidbina has quit [Ping timeout: 258 seconds]
<lovesegfault>
first release of 2020 :D
<gchristensen>
tracer41matrix[m: sometimes, the best way to help teach is to help somebody with their homework. regardless, just keep in mind the cutlure we have here :)
<lovesegfault>
gchristensen: Wait, what's the pending eval?
<tracer41matrix[m>
gchristensen: do your homework kidd :)))
tracer41matrix[m has left #nixos [requested by gchristensen (okay)]
* lovesegfault
laughs
tracer41matrix[m has joined #nixos
<gchristensen>
tracer41matrix[m: this channel has a different culture than many other IRC channels. you're welcome to be part of the community if you would like to be
<samueldr>
elitism has to be left at the door
<alexherbo2>
clever: it worked, thanks
<tracer41matrix[m>
gchristensen: thank you for explaining me the rules
<tracer41matrix[m>
Ignorantia legis non excusat id say 101
AstroBadger has quit [Quit: ZNC 1.6.5+deb1+deb9u2 - http://znc.in]
<{^_^}>
[nix] @edef1c opened pull request #3298 → passAsFile: leave out the hash prefix → https://git.io/JexyD
<tracer41matrix[m>
oof
<gchristensen>
as I said, it is up to you :)
<tracer41matrix[m>
this troll is so weak
<gchristensen>
I agree
<DigitalKiwi>
am i about to witness the first ban ever
leothrix has joined #nixos
<tracer41matrix[m>
Ignorantia legis non excusat id say 101
<tracer41matrix[m>
dude youre not playing with the rules