o1lo01ol1o has quit [Remote host closed the connection]
camsbury has quit [Remote host closed the connection]
bhipple has joined #nixos
camsbury has joined #nixos
random_yanek has joined #nixos
<NemesisD>
is there any concept of a strict mode or something that will error if i've provided an unused attr? like i'm trying to use haskellPackage.shellFor and i'm trying to override the ghc version, but i keep trying setting attributes that i think will override it but they get silently ignored
<symphorien>
a way to see if a= 1; is used is to replace it with a=assert false; 1
camsbury has quit [Ping timeout: 245 seconds]
<symphorien>
> { a = assert false; 1; b = 2; }.b
<{^_^}>
2
<symphorien>
> { a = assert false; 1; b = 2; }.a
<{^_^}>
assertion failed at (string):218:7
<NemesisD>
clever!
mkoenig has quit [Ping timeout: 245 seconds]
<dckc>
diagnostics from nix-build look really crummy in emacs compile mode. suggestions?
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #55465 → desktop-files-utils: add setupHook to remove mimeinfo.cache → https://git.io/fhHa9
<NemesisD>
is there a way to dump the dependencies? i'm setting ghc = pkgs.haskell.compiler.ghc822 and yet its saying Build with /nix/store/b39iibf198q06a0k4vq0lfzzmirq2hc3-ghc-8.6.3
ikitat has joined #nixos
<dckc>
,callPackage
<{^_^}>
If a Nix file foo.nix starts with something like `{ stdenv, cmake }:`, you can build it with `nix-build -E '(import <nixpkgs> {}).callPackage ./foo.nix {}'`
<srhb>
NemesisD: Normally you'd want to use haskell.packages.ghc822.shellFor
mkoenig has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
<srhb>
NemesisD: haskellPackages always refers to the latest stable package set (meaning, at the moment, ghc863 for unstable)
<{^_^}>
[nixpkgs] @worldofpeace pushed commit from @romildo to master « plano-theme: 3.30-2 -> 3.30-3 »: https://git.io/fhHad
<srhb>
dckc: callPackage provides it
<NemesisD>
srhb: so in my default.nix i use pkgs.haskell.packages.ghc822.extend to add my package. and in shell.nix i use pkgs.haskell.packages.ghc822.shellFor to instantiate it with dev tools? so then i guess my next step is to add a parameter to both of those files to take a ghc and default it to "ghc822"?
goibhniu has quit [Ping timeout: 246 seconds]
* dckc
scratches head
<dckc>
suppose I wanted to pass my own cmake
<srhb>
NemesisD: Maybe I should read up, I've no idea why you're using extend.
<srhb>
NemesisD: If you're doing that, your shell should be using the same hpkgs with shellFor
ng0 has quit [Quit: Alexa, when is the end of world?]
<qyliss>
dckc: I think you want to put the config stuff in the first attribute set
<NemesisD>
so i'm extending it to add my package as well as a couple of unreleased proprietary packages that are sitting in subdirs. also using dontCheck on some packages whose test suites don't pass although honestly i'd like to turn that off universally
<srhb>
NemesisD: I recommend that you take care to deal with only one package set across your project.
<NemesisD>
srhb: yeah i reference hpkgs from shell.nix
<dckc>
no joy
<srhb>
NemesisD: So, factor out a pkgs instantiation that you can share across them.
<srhb>
dckc: NixOS options are not nixpkgs options
slack1256 has joined #nixos
<srhb>
dckc: What are you trying to do?
<qyliss>
oh, whoops, I missed that
Rusty1 has joined #nixos
<dckc>
I'm trying to use jboss (well, wildfly) with nix installed on Ubuntu (or CentOS)
o1lo01ol1o has quit [Remote host closed the connection]
<dckc>
I'm trying to upgrade our i2b2 middle tier, and I'd rather not use ad-hoc methods as we've done in the past. the middle tier is (mostly) a jboss/wildfly web app
<srhb>
dckc: I recommend that you forgo all the NixOS components for now. Making the play with OtherOS is involved.
<dckc>
my desktop runs Ubuntu but the deployment platform is CentOS
<srhb>
dckc: And if you do want to do it anyway, you're probably better off building a custom systemd service builder for your purpose.
<srhb>
Rather than relying on the existing NixOS module.
<NemesisD>
srhb: is there a global way to disable 3rd party dependency tests and haddocks? ghc build times are long enough, lol
<dckc>
but I'm really confused about how scoping works... is callPackage magic? or is it just a function?
<srhb>
NemesisD: You can override the haskell package set's mkDerivation.
drakonis has quit [Quit: WeeChat 2.3]
<ottidmes>
dckc: just a function
<simpson>
dckc: In Nix, there's a few builtins for introspection. callPackage looks at the signature of its callee to figure out what to pass for defaults.
<dckc>
if I'm trying to put a square peg (nixos) in a round hole (nixpkgs) can anybody explain what the actual failure mode is? why is nix-build silently reporting success without doing anything?
<ottidmes>
dckc: there is a builtin called functionArgs, that allows you to check the attributes expected of an function expecting an attrset as an argument, i.e. it can determine that { stdenv, git }: ... will have "stdenv" and "git" as arguments
<bhipple>
Does ahnyone know what's going on here?
<dckc>
ah. so the magic is functionArgs
<simpson>
dckc: But to add on to what I've said, it would seem that the current incarnation of callPackage is a bit simpler than this. You could still write most of it yourself, if you had the Nix object model in your mind and were in enough of a self/super mood.
<dckc>
just a guess, bhipple : base16 vs base32
<srhb>
dckc: Re. silent "failure" -- what you're doing is essentially nix-build -E '(import <nixpkgs> {}).callPackage ({...}: {}) {}'
Guest52244 has quit [Ping timeout: 250 seconds]
<srhb>
Which in turn is pretty much nix-build -E '{}'
<srhb>
Which is both pretty silent and pretty boring.
<ottidmes>
bhipple: yep, like dckc mentions, most fetchers allow base16 and base32, the default tends to be base32, and I believe base64 is not also supported, at least in some of the latest Nix tools
<ottidmes>
s/not/now
<bhipple>
ottidmes: ah yes, you ad dckc are right; the longer one is in hex while the shorter one is indeed base32
<bhipple>
thanks. I notice a lot of the python packages happen to be in base16, though it looks like it gets converted to base32 for output error msgs when the prefetcher fails
<simpson>
bhipple: Likely pregenerated by pypi2nix or some other tool.
Melkor333 has quit [Ping timeout: 250 seconds]
<ottidmes>
bhipple: check out nix-hash to convert hashes
<ottidmes>
bhipple: yeah, hash mismatched right now are always reported in base32
<bhipple>
ottidmes: oh nice, I didn't know about nix-hash. Neat!
<bhipple>
ottidmes: that's a lot more convenient than googling for online converters in chrome
<srhb>
There's also nix to-base16, to-base32, to-base64
<ottidmes>
bhipple: another problem with that is that there are multiple base32 variants
<bhipple>
nix-hash --type sha256 --to-base32 <longer hash> produces the shorter hash, as expected
<bhipple>
Interestingly, it assumes md5 if no type is specified, which seems like a legacy default since md5 is generally no longer used
mounty has quit [Quit: Konversation terminated!]
<ottidmes>
bhipple: md5 was the default in the beginning, so yeah, legacy default
<ottidmes>
bhipple: its now deprecated though, best to use sha256
mkoenig has quit [Ping timeout: 250 seconds]
cydf has quit [Ping timeout: 240 seconds]
mounty has joined #nixos
o1lo01ol1o has joined #nixos
drakonis has joined #nixos
<{^_^}>
[nixpkgs] @markuskowa pushed to revert-53126-submit/megasync-3.7.1.0 « Revert "megasync: init at 3.7.1.0 (#53126)" »: https://git.io/fhHVG
<{^_^}>
[nixpkgs] @markuskowa opened pull request #55467 → Revert "megasync: init at 3.7.1.0" → https://git.io/fhHVZ
mkoenig has joined #nixos
<neonfuz>
so nix-env wants to update my openjdk-8 to openjdk-11
<neonfuz>
if I explicitly install jdk8 will it not do this?
<simpson>
Does bash go into POSIX sh mode when invoked as `sh`? Or does it require POSIXLY_CORRECT?
Maxdamantus has quit [Quit: brb]
<srhb>
simpson: It does go into one of the POSIX sh modes iirc.
<srhb>
Strangely it's not fully compatible
htoc^ has joined #nixos
<dckc>
those incompatibilities seem to be in dark corners I never hit, fortunately
rcshm has quit [Read error: Connection reset by peer]
<srhb>
Admittedly I seldom verify that the behaviour is POSIXly correct :P
eadwu has quit [Quit: WeeChat 2.3]
rcshm has joined #nixos
<dckc>
ok... so I've got wildfly starting and griping: Could not create server data directory: /nix/store/3avvma7bjb32bqz9qnywdlwmwkpkf6aw-wildfly/standalone/data
<dckc>
that preExec clue seems relevant
eadwu has joined #nixos
<simpson>
Progress!
<{^_^}>
[nixpkgs] @markuskowa pushed 0 commits to revert-53126-submit/megasync-3.7.1.0: https://git.io/fhHVW
<dckc>
gonna declare victory for the day. thanks, everybody
maximiliantagher has joined #nixos
Maxdamantus has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @worldofpeace pushed commit from @eadwu to master « vscode-extensions.ms-vscode.cpptools: 0.20.1 -> 0.21.0 »: https://git.io/fhHVF
marusich has quit [Ping timeout: 250 seconds]
ikitat has quit [Ping timeout: 272 seconds]
<kyren>
is nix-shell supposed to change the NIX_PATH variable, and if so can I prevent it from doing that?
eadwu has quit [Quit: WeeChat 2.3]
<kyren>
before using `nix-shell -p nix-prefetch-github` my NIX_PATH is 'nixpkgs=/etc/nixpkgs', after it looks to be reset to a default which looks at /nix/var/nix/profiles/per-user/root/channels/nixpkgs
<vonfry>
I have a pr for a unimportant tool several days ago and it is merged in master branch. I want to know what should I do let it merge into stable branch? Or just wait?
mizu_no_oto has quit [Quit: Computer has gone to sleep.]
nD5Xjz has quit [Ping timeout: 268 seconds]
<ottidmes>
vonfry: in less than 2 months 19.03 stable will be released, which most likely will include your tool as well, so its probably best to just wait rather than to try and get it backported to 18.09. In the meantime you can use nixos-unstable together with stable.
<ottidmes>
kyren: I don't think nix-shell does anything special for it, my NIX_PATH remains unchanged at least, so I guess it has to do with how/where you have set NIX_PATH
<kyren>
it must be, I set it in my "darwin-config" because I'm using nix-darwin... maybe it's a bug in nix-darwin?
eadwu has joined #nixos
silver has quit [Read error: Connection reset by peer]
<ottidmes>
kyren: could be, I have never used Darwin, let alone nix-darwin
<kyren>
yeah, it doesn't happen on at least one of my linux machines
vonfry has quit [Quit: WeeChat 2.3]
<ottidmes>
maybe ask on #nix-darwin? I would have expected many more people complaining about it, if this never worked properly before, so it might be that you are doing something differently than most, but that is my guess
<kyren>
oh, I didn't know there was a #nix-darwin, thanks
o1lo01ol1o has quit [Remote host closed the connection]
Edes has quit [Quit: Lost terminal]
eadwu has quit [Quit: WeeChat 2.3]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 240 seconds]
nD5Xjz has joined #nixos
petrkr has quit [Remote host closed the connection]
slack1256 has joined #nixos
marusich has quit [Remote host closed the connection]
<kyren>
I am not having a great day, it's like everywhere I step I step on bugs, I appear to be hitting https://github.com/NixOS/nix/issues/1697 OR I don't know what I'm doing and I can't tell the difference
<{^_^}>
nix#1697 (by knedlsepp, 1 year ago, closed): builtins.fetchGit fails when fetching new versions of a repository
slack1256 has quit [Remote host closed the connection]
<kyren>
I'm getting fatal: not a tree object\nerror: program 'git' failed with exit code 128 after using bultins.fetchGit a second time :/
<mdash>
ottidmes: i'm fooling with bitwarden_rs again, trying to add smtp config
<mdash>
ottidmes: was there a reason you put the env file in /etc instead of in the nix store?
noonien has joined #nixos
<noonien>
hello folks!
<noonien>
are there any examples of using make-disk-image?
<kyren>
okay, I think I'm starting to understand what my problem is, let me back up a bit, I'm trying to load nixpkgs from builtins.fetchGit, and I appear to be running into problems because nixpkgs does things like "loadModule <nixpkgs/nixos/modules/misc/assertions.nix> { }", and nixpkgs isn't a *directory*
<kyren>
is there a way to get nixpkgs from builtins.fetchGit that will work, what I'm trying to do is snapshot nixpkgs via the json output of something like nix-prefetch-git
<ottidmes>
mdash: preference I guess, I expect it to be there in other distros and dislike having to read the service files to pinpoint the env file in use, but the counter argument is probably that it pollutes /etc unnecessarily
<ottidmes>
mdash: when you say trying to add smtp config, do you mean use my config options for that, because my PR now includes them: https://github.com/NixOS/nixpkgs/pull/55413
<{^_^}>
#55413 (by msteen, 1 day ago, open): bitwarden_rs: init at 1.4.0
<kyren>
okay, I also understand what that fetchGit error message is, it's because I'm not specifying a ref parameter to builtins.fetchGit because I'm using the output of nix-prefetch-git which doesn't have it
pie___ has joined #nixos
ryantrinkle has quit [Ping timeout: 246 seconds]
<ottidmes>
noonien: you mean to make your own live CD or something else?
maximiliantagher has joined #nixos
<noonien>
i want to create an image to run with qemu, either a qcow, an .img or an .iso, doesn't really matter
<ottidmes>
mdash: any reason not to put the env in /etc?
<mdash>
ottidmes: I'm just used to seeing it as a store path
<ottidmes>
mdash: curses! so I went to all the trouble trying to package it for nothing
<mdash>
in general I only think about generated stuff living outside the store if things are hardcoded to look for it in a specific place
<mdash>
ottidmes: yeah i spent a couple hours on it too before i found this, heh
<ottidmes>
mdash: makes sense, I will put it in /nix/store
DrLambda has quit [Ping timeout: 250 seconds]
sinner has joined #nixos
<ottidmes>
mdash: can always create some shell functions to make it easy to get access to those paths, because there will be plenty of other services that don't have it in /etc, so my argument does not really hold
sinner is now known as Guest69325
jtojnar has quit [Ping timeout: 272 seconds]
ikitat has joined #nixos
<ottidmes>
mdash: I know its me finding an execuse to keep it (common practice of programmers after spending a lot of effort on a piece of code), but since the older version in the PR (due to Rocket 0.4 added to bitwarden_rs version 1.5 breaks with rustc 1.31) is build using the Web Vault 2.4.0 and that fork only starts at 2.5.0, I am keeping it for versions older than 2.5.0
rcshm has quit []
ikitat has quit [Ping timeout: 244 seconds]
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mdash>
ottidmes: Sure.
daniele- has joined #nixos
<mdash>
ottidmes: I gave up and used the overlay for nightly, since I'm not doing an upstream PR :)
<ottidmes>
mdash: once staging merges, it will have rustc 1.32 that might be able to compile it, version 1.33 at least does, but I figured, I can always make a new PR when that happens, better to already have it in, and I know this version works, since I have been using it for months
maximiliantagher has quit [Ping timeout: 246 seconds]
renais has quit [Ping timeout: 250 seconds]
praetorg_ has joined #nixos
drakonis has quit [Quit: WeeChat 2.3]
kvda has joined #nixos
_praetorg_ has quit [Ping timeout: 250 seconds]
qualiaqq has joined #nixos
renais has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bhipple has quit [Remote host closed the connection]
endformationage has quit [Quit: WeeChat 2.3]
<mdash>
ottidmes: definitely
<mdash>
ottidmes: I look forward to having it in nixpkgs, let me know if I can help with anything
renais has quit [Ping timeout: 246 seconds]
<ottidmes>
mdash: just using and testing the PR as it is right now would help. I have only tested the basics, my server is still running my old package/module, but I cannot update it right now. Like you mentioned SMTP, does that work with the module as it is right now? I never tried it
ddellacosta has quit [Ping timeout: 246 seconds]
renais has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
renais has quit [Ping timeout: 246 seconds]
renais has joined #nixos
nD5Xjz has quit [Ping timeout: 250 seconds]
palo1 has joined #nixos
renais_ has joined #nixos
renais has quit [Ping timeout: 268 seconds]
palo1 is now known as palo
nD5Xjz has joined #nixos
pointfourone has joined #nixos
pointfourone has quit [Client Quit]
o1lo01ol1o has joined #nixos
maximiliantagher has joined #nixos
o1lo01ol1o has quit [Ping timeout: 246 seconds]
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #nixos
maximiliantagher has quit [Ping timeout: 246 seconds]
leotaku has joined #nixos
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #nixos
yayforj has joined #nixos
<leotaku>
Good morning. I've just experienced an issue where, because of a "bad" overlay I had nix-build fill up all available memory and essentially lock my system. Is that something I should report as a bug, or is that "expected" with certain types of wrongly configured overlays.
<leotaku>
*?
jtojnar has quit [Read error: Connection reset by peer]
jtojnar has joined #nixos
Maxdamantus has quit [Quit: brb]
Maxdamantus has joined #nixos
nD5Xjz has quit [Ping timeout: 245 seconds]
<ottidmes>
leotaku: if you think this could have been detected decidably somehow or is clearly a bug, only then I would make an issue for it
<ottidmes>
leotaku: if its just a case of bad code, then its not worth an issue, there are many ways to reproduce your issue that I would not consider bugs of Nixpkgs/Nix
DrLambda has joined #nixos
renais_ has quit [Ping timeout: 245 seconds]
<leotaku>
ottidmes: Interesting. But aren't there usually restrictions (eg. max recursion depth) so that the process maybe fails but doesn't crash the users machine?
nD5Xjz has joined #nixos
<ottidmes>
leotaku: there are some in place (you see this with infinite recursion being reported), but not all cases can be guarded against, I am not familiar enough with the implementation to know if there exists a max recursion depth, but in all liklihood your memory will already be full before this triggers, its a hard problem
yayforj has quit [Remote host closed the connection]
<bpye>
Are there any tools like NixOps for managing a NixOS server when not on a NixOS machine?
mobile_c has joined #nixos
<bpye>
The only thing I can think of right now is to set up some sort of container in which I'd run the NixOps deployment, but given that NixOps does most of the deployment process locally that sounds like it's gonna suck
<bpye>
By not NixOS machine, I mean a machine without Nix, be it a Windows machine, or something running *BSD... MacOS and Linux support are great but sadly I can't be in those environments always
<bpye>
Haha, as I said, sometimes I am on a Windows machine though, and try as I might, Nix runs like crap on WSL -> I'm well aware that's a WSL issue, it's filesystem perf leaves a lot to be desired
<bpye>
I can always run in a VM but then life sucks for other reasons
DrLambda has quit [Ping timeout: 250 seconds]
<clever>
bpye: i have a central dedicated machine i always run nixops on, and ssh to that
<mobile_c>
0.0 awesome
<bpye>
clever: How do you make editing files remotely suck less?
<clever>
bpye: i always use vim, so i just run vim under ssh
<bpye>
I have never found a way to make that work nicely, it's always been painful
<bpye>
Ah I see
<wedens>
bpye: if you use emacs, you can use TRAMP
<mobile_c>
does NixOS have support for m68k cross compilers
<clever>
bpye: ive also made a vim.nix that includes all of my standard config and plugins, so i just `nix-build vim.nix -o vim` and `./vim/bin/vi` and it works 100% like the local one
<bpye>
I can get by with vim, I've come to quite like vs code though which afaik lacks any decent remote editing solution
<clever>
bpye: samba is also an (ugly) solution, just map the remote disk to the windows box
<bpye>
The best I've ever found is an extension to support rmate - I've used samba in the past but it tends to screw up permissions
<wedens>
is there something like sshfs on windows?
<bpye>
Ah... There totally is an sshfs FUSEish driver for windows
<bpye>
I forgot that existed - it's been a couple years since I last had this problem
<wedens>
doesn't seem like this module writes any config files
<leotaku>
cswl: Also with services specified in configuration.nix you are supposed to only configure them through their options. You wont be able to edit the configuration files on the filesystem.
<balsoft>
Yeah
<balsoft>
nixpkgs doesn't store any configuration
<balsoft>
It's exactly the job of NixOS
<balsoft>
nixpkgs does the opposite - make the same packages behave the same way everywhere
<balsoft>
Config file would break that
maximiliantagher has joined #nixos
<cswl>
Where does it save systemd unit files
<balsoft>
Nowhere
<balsoft>
NixOS generates them
<balsoft>
nixpkgs is free of any systemd reference
<balsoft>
s
<balsoft>
It does have systemd as a package, that's about it.
<balsoft>
It has other inits as packages, but NixOS only uses systemd (so far)
<leotaku>
balsoft, cswl: Well it stores them in the nixpkgs store.
<balsoft>
It's not nixpkgs store!
<balsoft>
It's nix store.
<balsoft>
If you want nix to generate configuration files and startup scripts, use nixos
<leotaku>
balsoft: Sorry, I messed up the terminology.
maximiliantagher has quit [Ping timeout: 268 seconds]
<balsoft>
Are there any standalone WM users here?
<balsoft>
I mean no DE
<leotaku>
cswl: Are you on NixOS or just using the Nix package manager?
<balsoft>
I need some help with Qt icon theme
<balsoft>
I can't seem to get any help anywhere. Tried stackoverflow, reddit/r/linuxquestions, #qt
<balsoft>
The problem is that Qt doesn't recognize KDE icon theme
<balsoft>
Set in kdeglobals
<balsoft>
It did recognize it earlier
zupo_ has joined #nixos
<leotaku>
balsoft: I am using FVWM currently, but I don't know how to help you...
zupo has quit [Ping timeout: 250 seconds]
<cswl>
Just using NIx package manager
<leotaku>
cswl: Well then all you can do is install packages using eg. nix-env nix-shell. You won't be able to manage your configurations using only Nix.
<goibhniu>
hi balsoft, can you provide more details, an example of what used to work and no longer works? Do you see any error message when launching a Qt application?
<cswl>
Well.. there must be some way to change configuration options :/
<leotaku>
cswl: as balsoft said, thats what NixOS is for.
<{^_^}>
[nixpkgs] @vcunat pushed 282 commits to staging-next: https://git.io/fhHoy
_kwstas has quit [Remote host closed the connection]
<goibhniu>
balsoft: also, does running `kbuildsycoca5` help?
<cswl>
It doesnt seem to read this..
<cswl>
The system configuration is located at /etc/lxc/lxc.conf or ~/.config/lxc/lxc.conf for unprivileged containers.
<goibhniu>
oh, you probably don't have that without KDE/plasma
<leotaku>
cswl: The applications installed using Nix should just read their config files from the usual locations. I would be surprised if this was a Nix/Nixpkgs issue.
<kyren>
I'm not sure this is a nixos question exactly, but I have a cheapo dedicated server with a bunch of containers on it, and the one that has gitlab in it is *very slow* to restart, and this constantly causes headaches for me because when I do a nixos-rebuild it always causes "error(s) occurred while switching to the new configuration", and I believe this is just timing out
<kyren>
Is there a way to increase this timeout
<kyren>
seriously, starting the container that has gitlab in it takes a good 45 seconds
monotux has joined #nixos
i1nfusion has quit [Remote host closed the connection]
simukis has quit [Quit: simukis]
i1nfusion has joined #nixos
simukis has joined #nixos
qualiaqq has quit [Quit: ERC (IRC client for Emacs 26.1)]
shibboleth has joined #nixos
<leotaku>
kyren: I think such errors should be logged by systemd. Can you maybe try "journalctl -xe" to confirm that the problem actually is what you think it is?
<slyfox>
'nix-channel --update' fails with an obscure failure for me: "error: while setting up the build environment: getting attributes of path '': No such file or directory" http://dpaste.com/2QRRPCH.txt
<slyfox>
Where does that '' path come from?
<balsoft>
I can provide some before/after scrots of trojita
<balsoft>
And I can provide exact versions of nixpkgs that work and don't work
<balsoft>
I have a feeling it's not nixpkgs issue per se, but an issue with Qt
<balsoft>
But I couldn't find any commits in Qt source that would change Icon behaviour in 5.12
dermetfan has quit [Ping timeout: 245 seconds]
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
ikitat has quit [Ping timeout: 268 seconds]
maximiliantagher has joined #nixos
<goibhniu>
balsoft: it might be worth testing with a fresh user account
<goibhniu>
just in case Qt is caching something store path that no longer exists
<goibhniu>
^something^some
elgoosy has joined #nixos
Ariakenom has joined #nixos
<kyren>
so, okay given what I have now figured out, I suppose my question is: is there a way to control the TimeoutStartUSec service option for a nixos container, because apparently one of mine likes to take longer than *looks at current value of option* 1 minute and 30 seconds
nD5Xjz has quit [Ping timeout: 272 seconds]
<ottidmes>
slyfox: let me guess, Nix install, i.e. not NixOS? and on Arch?
<slyfox>
it's not a nixos and gentoo
<ottidmes>
slyfox: would you mind creating an issue for this over at https://github.com/NixOS/nix/issues, I have been seeing people running into this problem multiple times now, yet I don't see an issue created for it
<ottidmes>
I mean at least 7 people encountering the problem and asking about it warrants an issue
<{^_^}>
nix#2673 (by trofi, 12 seconds ago, open): 'nix-channel --update' stopped working on nix-2.2.1: error: while setting up the build environment: getting attributes of path '': No such file or directory
rprije has quit [Ping timeout: 245 seconds]
<slyfox>
Looks like this was enough to repair it: nix-build '<nixpkgs>' -A bash && sudo systemctl restart nix-daemon
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @r-ryantm to master « python37Packages.httmock: 1.2.6 -> 1.3.0 (#55195) »: https://git.io/fhH6n
<leotaku>
kyren: Sorry but I've not been able to find any configureable option to enable that functionality. You of course would be able to change the option by editing the nixpkgs source, but thats undesireable.
<ottidmes>
slyfox: thank you! hopefully this helps getting it fixed, or at least we have something to point to if someone else runs into the issue :)
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @r-ryantm to master « python37Packages.libcloud: 2.3.0 -> 2.4.0 (#55219) »: https://git.io/fhH6W
<ottidmes>
leotaku: do you know the file, I tried searching to see if I could think of a way to configure it, but without any source code I am at a loss to help
knupfer has quit [Remote host closed the connection]
<leotaku>
kyren: Maybe you could just set "autoStart" to false and then manually restart the services?
<kyren>
I feel like it's a bit silly, since the container systemd service is kind of "composed" of potentially many, many sub-services, so I feel like there ought to be a solution here where a timeout only occurs if the services *inside* the container time out
<kyren>
leotaku: yeah I suppose I could do that, I've been actually just manually restarting the containers anyway since if they get timeout killed they're in a weird transition state
<kyren>
ottidmes: that works perfectly, thank you very much! (also now I know how to map both the key and value of an attribute set to make a new attribute set)
b has joined #nixos
<kyren>
leotaku: the nixos module system is great, it will merge things together intelligently even if they can't be combined, if two modules have the same config with the same non-mergable value (like a boolean) it will successfully merge, but if they conflict it won't
Thra11 has joined #nixos
pie_ has joined #nixos
<immae>
Hey there! Since my last nix upgrade (2.1.3 -> 2.2.1) the nix repl began to be really hard to use (no completion, strange behavior of keys, etc.) It seems to be readline-related, but did anything change in that regards in nix?
maximiliantagher has joined #nixos
jomik has joined #nixos
<immae>
(nb: I tried downgrading back to 2.1.3, all other things unchanged, and it works as expected, so it seems to come from nix itself)
maximiliantagher has quit [Ping timeout: 250 seconds]
<tilpner>
nix@de599733 looks relevant ("repl: give user the choice between libeditline and libreadline"), but it's from Nov. 15
<tilpner>
To get my exact version, nix-store -r /nix/store/bqipf41ygbhkppc469fz9x18yxyr0hnx-nix-2.2
<tilpner>
(Or nix-build as usual)
andi- has joined #nixos
silver has joined #nixos
<immae>
Yes it works
<immae>
So it’s export EDITLINE_CFLAGS="-DREADLINE"; export EDITLINE_LIBS="/usr/lib/libreadline.so" that blew the thing, but I cannot find libeditline in my distribution :/
<tilpner>
,locate libeditline.so.1
<{^_^}>
Found in packages: editline
<tilpner>
Oh, hmm
<immae>
Yes, it’s not a nix problem here, I’ll find my way don’t worry :)
<infinee>
oh my first encounter with a package that isn't latest. I need a newer version of nixos.delve. Can someone point me to what I can do to get the newer version of delve included in nixpkgs?
<srhb>
infinee: Find the expression in the nixpkgs git repo, bump the version and the hash, try building it, if it works PR it :)
<infinee>
srhb: bump it locally?
<infinee>
the nixpkgs git repo?
<srhb>
infinee: Yeah.
<infinee>
I've not installed nixpkgs yet. All my stuff has been installed via nixos
<srhb>
infinee: You can just clone the repo :) There's no "install" needed until you want to pull in the package locally.
<srhb>
Which can be done through a variety of means, one being overlays.
<srhb>
(They are documented in the nixpkgs manual)
maximiliantagher has quit [Ping timeout: 246 seconds]
sicklork1n has joined #nixos
bpa has quit [Remote host closed the connection]
xkapastel has quit [Quit: Connection closed for inactivity]
bpa has joined #nixos
orivej has joined #nixos
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<infinee>
Strange, nix-env -qa delve shows 1.0.0 but the nixpkgs repo I cloned has 1.1.0
<infinee>
and nix-channel --update didn't change anything
renais_ has joined #nixos
renais has quit [Ping timeout: 250 seconds]
<ottidmes>
infinee: not strange at all, how is nix-env to know about your cloned repo, try this instead: nix-env -f /path/to/nixpkgs/checkout -qa delve
<infinee>
ottidmes: I've done that. I've also done the 'nix-build -A delve' and I see multiple versions in /nix/store. However my nix-shell still picks up old one
o1lo01ol1o has joined #nixos
<ivegotasthma>
is there a way to orchestrate services with bare nix?
<symphorien>
the easy way is to pass -I nixpkgs=/path/to/repo to all nix commands you want to use the clonse
<{^_^}>
[nixpkgs] @dotlambda pushed to master « nixos/home-assistant: make config.http.server_port an integer »: https://git.io/fhHPs
<simpson>
ivegotasthma: There's a couple tools that are incidentally basically bare Nix for orchestration, like https://github.com/xtruder/kubenix
<simpson>
But Nix can't do the impure work of changing a distributed system's state all on its own.
<ivegotasthma>
I'm aiming for a single system
<ivegotasthma>
right now
<ivegotasthma>
I want to start a postgres process, nginx and a python process
o1lo01ol1o has quit [Ping timeout: 250 seconds]
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<simpson>
NixOS can do that on its own.
<ivegotasthma>
it's not a nixos system
<ivegotasthma>
there's a debian9 server I want to install nix on
o1lo01ol1o has joined #nixos
<simpson>
And it has to remian Debian? Then there's not much you can do besides the obvious.
o1lo01ol1o has quit [Ping timeout: 246 seconds]
o1lo01ol1o has joined #nixos
maximiliantagher has joined #nixos
balsoft has joined #nixos
Jetien_ has joined #nixos
balsoft has quit [Remote host closed the connection]
lsyoyom has quit [Ping timeout: 250 seconds]
maximiliantagher has quit [Ping timeout: 268 seconds]
fusion809 has joined #nixos
rauno has quit [Remote host closed the connection]
Jetien_ has quit [Ping timeout: 250 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
<steveeJ>
is there any way to get the fusermount wrapper in a nixos-container?
teehemka_ has joined #nixos
<steveeJ>
oh it's already there, just not found by the systemd service
lsyoyom has joined #nixos
<infinisil>
,dnw jomik
<{^_^}>
jomik: "Does not work" isn't very helpful: What doesn't work? What's the error?
drakonis has joined #nixos
sicklork1n has quit [Ping timeout: 268 seconds]
<steveeJ>
dang, of course there's no /dev/fuse in the container
<steveeJ>
declaring a bind-mount for it on the conatiner doesn't help much. "mount helper error: fusermount: failed to open /dev/fuse: Operation not permitted". am I the first one to try this?
MinceR has quit [Ping timeout: 268 seconds]
<infinisil>
,dnw = "Does not work" isn't very helpful: What specifically doesn't work? What command did you run? What's the error?
<{^_^}>
dnw redefined, was defined as "Does not work" isn't very helpful: What doesn't work? What's the error?
<steveeJ>
oh! there is "containers.<name>.allowedDevices"
MinceR has joined #nixos
<{^_^}>
[nixpkgs] @etu opened pull request #55483 → Created a folder for radio-related applications → https://git.io/fhHPA
alex_giusi_tiri has joined #nixos
ashkitten has quit [Ping timeout: 250 seconds]
<kumikumi>
What is currently the easiest/recommended way to create and run LXC containers under NixOS? I'm looking to run some kind of Ubuntu
DrLambda has quit [Ping timeout: 240 seconds]
<lassulus>
virtualisation.lxd.enable = true;
<lassulus>
and add yourself to the lxd group
<jomik>
infinisil: I wish I could say - it simply does not show anything there. And I have no clue how to debug it.
<jomik>
There is like, no output in any of the services that indicate an error.
<kumikumi>
lassulus: thanks
<lassulus>
kumikumi: then you can just run lxc launch ubuntu:16.04 blabla
<infinisil>
jomik: Yeah, okay but what doesn't work actually?
<kumikumi>
lassulus: and it works as a non-root user as long as the user is in the ldx group?
<lassulus>
yes
<kumikumi>
okay great
mmlb3 has quit [Ping timeout: 246 seconds]
<jomik>
So, when I add tray to my endWidgets, I'd expect a tray to show up with nm-applet and flameshot. But nothing appears there.
<jomik>
I have the other taffybar widgets though.
<lassulus>
ah, maybe you will also have to run lxc init
<infinisil>
jomik: So taffybar starts no problem?
<jomik>
Yeah, no errors in taffybar's logs and it runs "fine"
<jomik>
Just sni-tray doesn't seem to pick up any icons/widgets or what you want to call them.
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fhHXk
acarrico has joined #nixos
<ottidmes>
infinisil: how many PRs do you think fall under (!bot && changed_only_version_and_hash && triggered_rebuilds == 1), because those should be relatively safe to fast track being committed to master
ashkitten has joined #nixos
<infinisil>
jomik: Yeah I don't know anything about trays, am never using them
<infinisil>
ottidmes: No idea, but that would be interesting
Thra11 has quit [Ping timeout: 250 seconds]
erictapen has quit [Ping timeout: 245 seconds]
erictapen has joined #nixos
<jonge>
i made a fresh install of nix yesterday on my mac (2.2.1) and nix-channel --update and nix-env -i... always fails with "error: unexpected end-of-file". what can that be? i find no real help online since that happened to people in the past, but in different situations and their fixes don't apply for me.
alex`` has joined #nixos
<thefloweringash>
Maybe nix#2523 ? Did you try the `OBJC_DISABLE_INITIALIZE_FORK_SAFETY` workaround?
<noonien>
how can i the path to where a channel files are?
init_6 has joined #nixos
renais_ has joined #nixos
teehemka_ is now known as teehemkay[away]
teehemkay[away] has quit [Quit: ZZZzzz…]
<pbb>
Actually loading amdgpu crashes the entire system..
<fpletz>
pbb: I know this sounds weird but are you boot via uefi or legacy mode?
<teehemkay>
Hi everyone!
<fpletz>
pbb: I've a lenovo A485 and have the same problem with legacy boot, uefi works
<pbb>
I am using uefi with systemd-boot
<pbb>
And it's a Ryzen Desktop system with a Vega 56 GPU
<teehemkay>
I'm doing a fresh install of Nix on MacOS Mojave 10.14.3 in multi-use mode.
<teehemkay>
The installation completes successfully but when I try to verify the install with `$ nix-shell -p nix-info --run "nix-info -m"` I get get an error: "error: unexpected end-of-file". I get the same error whenever I try to install anything now.
<teehemkay>
Any idea what's happening? I search the interweb but the only reference to the error on macOS points to an issue that has been resolved.
<pbb>
fpletz I just discovered that I had not ran nix-channel --update so it was using latest linuxPackages from 18.09. With 4.20 it works \o/
<NemesisD>
hi all, i've noticed a pattern of { pkgs ? import <nixpkgs> {}}: at the beginning of nix expressions to optionally pass in a fixed package set. is it sensible to pin a specific version in shell.nix and then have default.nix and others take the argument?
yayforj has quit [Ping timeout: 272 seconds]
<cswl>
How do I add a system package if im using Nixpkgs only
<tilpner>
AFAIK there is no notion of system packages for Nix-on-not-NixOS, but you can check your PATH for more certainty
<tilpner>
You can probably do that yourself, add a system profile and add its PATH to some global config file (/etc/profile?)
rcshm has quit [Remote host closed the connection]
<tilpner>
jluttine: nix-build --no-out-link '<nixpkgs>' -A hello
* tilpner
recommends alias nip='nix-build --no-out-link "<nixpkgs>" -A'
<gchristensen>
nip?
<jluttine>
tilpner: nice, thanks!!
rcshm has joined #nixos
<ottidmes>
infinisil: I did some querying and found that 559 PRs are rebuilding < 10 packages on Linux, 457 of those are mergable, and only 186 of those are updates (they adhere to the "package: ... -> ..." title format)
Glider_IRC__ has quit [Remote host closed the connection]
<ottidmes>
gchristensen: "NIx Path" I guess
dermetfan has quit [Ping timeout: 240 seconds]
<tilpner>
gchristensen: All my Nix related aliases start with ni, all NixOS related ones with no, and then a letter for whatever action it does
<tilpner>
"NIx Package" works too, but it's arbitrary of course
ddellacosta has joined #nixos
* tilpner
also has snors for sudo nixos-rebuild switch
<ottidmes>
lol, I had some of those aliases, but if I did not use them frequently I would totally forget about the name, I just have nixos-rebuild alias to sudo nixos-rebuild
<tilpner>
I use them often enough, and delete old aliases that I tend to not use
<jluttine>
for now, i like to use the correct full commands so that i'll learn things. but good idea to use aliases once one is familiar with how to do things, just bored of typing them all the time
<ottidmes>
tilpner: I actually have a command called "nux" (Nix User eXperience) is my collection of helpful Nix related shell functions, all documented with a help and I added auto complete that gives me a nice list of the available commands
<tilpner>
That sounds a lot more fancy than my half-a-screen of aliases
* tilpner
is still looking for good pin management tooling
<ottidmes>
tilpner: pin management? passwords?
sinner has joined #nixos
sinner is now known as Guest64117
<tilpner>
No, to pin sources, so that you can put them into git. Right now I just have a script that writes JSON into a file, which is then passed to fetchTarball
<tilpner>
Like a more declarative nix-channel
rcshm_ has joined #nixos
rcshm has quit [Read error: Connection reset by peer]
Glider_IRC has quit [Remote host closed the connection]
<tilpner>
Oh, hadn't seen those before
<sicklorkin>
ottidmes: Oh..thanks for sharing those.. :)
<ottidmes>
I am cleaning up nix-prefetch at the moment, which will break nix-update-fetch's way of using it, but I plan to release 0.1.0 of both tonight or tomorrow
rcshm_ has quit [Read error: Connection reset by peer]
<sicklorkin>
anyone else have anything for show-and-tell?
rcshm has joined #nixos
<das_j>
Ok, weird problem. I'm building Android with nix and the build failed because my disk is full. So I'd like to perform a GC, but I'd like to keep the source for this time. What do I do? Will the source not be gc'ed when I have a nix-shell open?
<ottidmes>
the breakage will be limited to removing --diff and --with-position that were basically only there to make nix-update-fetch work, now I will be going about it differently requiring no hardcoded support in nix-prefetch for nix-update-fetch
<tilpner>
das_j: You want to explicitly create root for the source
<das_j>
tilpner: with --add-root?
<tilpner>
das_j: Where do you see add-root? I would have tried nix-builds --out-link
<tilpner>
(And are you actually building AOSP with Nix? That's scary)
<sicklorkin>
das_j: `--indirect", "--add-root`
<tilpner>
all-packages.nix is hopelessly out of order :(
<das_j>
tilpner: Yes, a friend of mine wrote some tools to build Android
<das_j>
it works right now but I'd like to rebuild it
<ottidmes>
tilpner: yeah, I am itching to clean all-packages.nix up, sorting it automatically and such, but I doubt I am allowed to do so
asymmetric has joined #nixos
<tilpner>
ottidmes: You just need to convince someone with push rights c.c
<tilpner>
Or per-categories indices, that might work too
<tilpner>
But then you're doing a major refactoring, and might as well do all sorts of other things, and nobody can decide what those are exactly (e.g. tagging)
<ottidmes>
would be cool if it was some script that needs to be run when you make a PR, that way it keeps being nice and start getting unsorted/inconsistent the moment the next PR gets in
<ottidmes>
biggest problem is getting the first big cleanup in, because wouldnt that break all other PRs changing the file?
pie_ has joined #nixos
<samueldr>
that's part of it
<samueldr>
one big issue with "machine re-ordering" is that it's not line-based
<samueldr>
so you need an AST and to build the code in a way that keeps the comments relevant
<samueldr>
which will be... hard
<samueldr>
some things are bunched up by topic
<ajs124>
tilpner, depending on what state we get it in and if you'll be there, you might get to see a lightning talk on building AOSP with nix :P
<samueldr>
ajs124: amazing <3
<das_j>
Spoiler: It's ugly
<samueldr>
there'll be a write-up too I hope?
<tilpner>
ajs124: If "there" is a physical location, odds are I'll miss it. But you'll make a video, right?
<samueldr>
(a video is fine too)
daniele- has joined #nixos
<sicklorkin>
I have a question about the installer. What\s the motivation behind having the user set the file mode bits rather than the installer itself? I mean the installer already sets the ownership (for multiuser anyway).
<sicklorkin>
I think this was a conscience decision, but I can't find the rational
<symphorien>
xss-lock launches i3lock automatically when the lid closes
<etu>
oh
<sicklorkin>
cswl: these may be completely unrelated issues then
<cswl>
Eh its working now..
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
rcshm has quit [Ping timeout: 240 seconds]
Izorkin has quit [Ping timeout: 268 seconds]
<kumikumi>
I'm having trouble with port forwarding from NixOS host to Ubuntu LXC guest. Tried following instructions at https://discuss.linuxcontainers.org/t/forward-port-80-and-443-from-wan-to-container/2042 . "curl localhost" inside container works (outputs a bunch of html), but outside the container I get: curl: (56) Recv failure: Connection reset by peer
<kumikumi>
I was wondering if there is something NixOS-specific I should be thinking about
eadwu has joined #nixos
Izorkin has joined #nixos
<kumikumi>
(also my use case is actually a game server, not a web server, web server is just easier to test/debug with)
<freusque>
Hello! Has anyone run into the situation where a script that uses a nix-shell shebang fails to run when called from a nix-shell environment?
<freusque>
The script runs fine when not called from a nix-shell and has the flowing shebang:
<mdash>
dunk: try 'nix search netcat', it should probably work :)
<mdash>
nix-index is also pretty useful, for searching by filenames
<cdyson37>
Hi. On editing /etc/nixos/configuration.nix and setting mutableUsers=false and adding a new user, then running 'nixos-rebuild switch', I can see that /etc/passwd has been modified, but the user I added is not present. Any idea what I've done wrong?
<freusque>
dunk: nix 2 is a recent rewrite of the nix CLI tools. if you have `nix` in your path, your system has nix 2.
<dunk>
Right
asymmetric has quit [Remote host closed the connection]
<dunk>
I'm using the nix docker image at the moment
<dunk>
Seems to work
zupo_ has joined #nixos
zupo has quit [Ping timeout: 245 seconds]
<dunk>
So, 'nix search netcat' seems to have worked
<ajs124>
Yes, someone please look at those. MariaDB is way out of date.
<dunk>
freusque: I am afraid I am pretty new to this. I was reading through the main docs, but then hit upon the netcat not installing thing. Is there a different set of docs I should look at?
<dunk>
I know that the -i flag is "install", but the -iA is opaque to me. And I also don't understand why I should prefix netcat with "nixpkgs", so I think there is some reading for me to do?
<freusque>
dunk afaik the proper approach is man, --help, roll with it
o1lo01ol1o has joined #nixos
<dunk>
kk
<freusque>
:)
justanotheruser has quit [Quit: WeeChat 2.2]
rfold has joined #nixos
<dunk>
Only problem is that the nixos Docker image has the man pages stripped out and I can't work out how to install them because... I can read the man pages
justanotheruser has joined #nixos
<dunk>
I assume that it's a trick like the netcat one above?
shabius has quit [Quit: Leaving]
<tilpner>
cdyson37: If you look at /run/current-system/activate, you'll find *-users-groups.json. With 'jq . < hash-users-groups.json' you can pretty-print it and verify that it looks right
<jabranham>
dunk: nix-env -A uses an attribute path, which I like to think of as a channel. So nix-env -A nixpkgs.netcat says "install netcat from the nixpkgs channel".
<dunk>
jabranham: How do I install the man pages?
<dunk>
:-D
<freusque>
the main docs are invaluable too don't get me wrong, but many things aren't in there, yet. you can have a look at the nixpkgs contributor guide too
<jabranham>
you could add another channel and call it "foo" and do nix-env -A foo.netcat to get foo's netcat
<dunk>
jabranham: I see
<jabranham>
dunk: no idea about the man pages, I don't use docker
<tilpner>
cdyson37: I don't what happens on UID collision. Also check if other users have the same uid
<tilpner>
*don't know
o1lo01ol1o has quit [Remote host closed the connection]
shabius has joined #nixos
shibboleth has quit [Quit: shibboleth]
<cdyson37>
Thank you! I'm an idiot - typo'd username :(
<{^_^}>
[nix] @LnL7 opened pull request #2674 → nix-daemon: add variable to disable fork safety → https://git.io/fhHyq
JosW has joined #nixos
<{^_^}>
[nixpkgs] @asymmetric opened pull request #55491 → nixos/xautolock: improve doc of time parameer → https://git.io/fhHym
o1lo01ol1o has joined #nixos
daniele- has quit [Quit: daniele-]
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @bhipple opened pull request #55492 → pythonPackages.testpath: build from src instead of pre-built binary → https://git.io/fhHyn
rcshm has quit [Remote host closed the connection]
<symphorien>
yes but what for
<symphorien>
,libraries rain1
<{^_^}>
rain1: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
_kwstas has joined #nixos
lsyoyom has quit [Ping timeout: 268 seconds]
rcshm has joined #nixos
<rain1>
$ nix-shell -p autoreconfHook -p libX11
<rain1>
error: undefined variable 'libX11' at (string):1:111
maximiliantagher has joined #nixos
<symphorien>
,locate libX11.so
<{^_^}>
Found in packages: xlibs.libX11
Ariakenom has joined #nixos
<symphorien>
you were close
Ariakenom has quit [Read error: Connection reset by peer]
simukis has left #nixos [#nixos]
Ariakenom has joined #nixos
maximiliantagher has quit [Ping timeout: 250 seconds]
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
simukis has joined #nixos
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
mbrgm has joined #nixos
Cale has joined #nixos
lsyoyom has joined #nixos
zupo has joined #nixos
<JosW>
Hi, I am a bit lost
<JosW>
Say i want to enable nginx, what are the steps to follow
<JosW>
do i add nginx in the pkgs list in configuration.nix?
<tilpner>
JosW: Do "man configuration.nix", then search for "services.nginx"
<{^_^}>
[nixpkgs] @etu merged pull request #55491 → nixos/xautolock: improve doc of time parameer → https://git.io/fhHym
<ottidmes>
symphorien: question about the bitwarden_rs PR, how will DynamicUser work with the backup service, if I supply the same User/Group as the other service I get access to the data is what you said, right? if so, then I can read the data, but how do I get the data written with the right backup user/group? All I can think of at the moment is using root, did you had something else in mind?
<symphorien>
ah msteen == ottidmes
<symphorien>
honestly I don't really know what bitwarden is
<ottidmes>
a password manager
<ottidmes>
but that should be irrelevant to the systemd setup
<symphorien>
what use case did you have in mind with this backup thing ?
<ottidmes>
if I accidentally wipe my hard drive, I dont want to lose all my passwords
camsbury has quit [Remote host closed the connection]
<symphorien>
a regular backup of /var/lib will do the trick
<ottidmes>
which is what this service does
<symphorien>
without anything specific to bitwarden
<symphorien>
the folder the service will see as /var/lib/bitwarden is in fact /var/lib/.somehash
<symphorien>
so if you backup all of /var/lib then fine
alex`` has quit [Quit: WeeChat 2.3]
spear2 has joined #nixos
<ottidmes>
symphorien: personally I dont backup /var/lib but only backup per application, and I don't think its always safe to just blindly copy over state
<ottidmes>
symphorien: if you consider it out of scope, I can get behind that, then I will just keep that bit for myself
camsbury has quit [Ping timeout: 246 seconds]
<symphorien>
ah no sorry, the directory seems to be /var/lib/private/bitwarden and owned by root
balsoft has quit [Ping timeout: 240 seconds]
<ottidmes>
symphorien: yeah, read about that, where /var/lib/private acts like a boundary
<symphorien>
I don't know what the better solution is
<symphorien>
I do feel that a specific backup service per app is too much, and that an option to choose where favicons are store might be superfluous, but well this is an opinion, and if you feel otherwise, please do so
<ottidmes>
well I assume sqlite has a backup commando for a reason, rather than just copying over the file, probably to prevent corruptions, so I consider being explicit about the backup the better solution, but whether it fits with the module or whether its up to you yourself to deal with that, thats the question (of course its all behind options anyway, but still)
lsyoyom has quit [Ping timeout: 250 seconds]
<symphorien>
makes sense
<ottidmes>
and this is basically what the author of the application suggest you do
<symphorien>
in that case, it seems dynamic users only get in the way.
<ottidmes>
at the moment, rather than simplifying things, it does make things more complicated yeah
<ottidmes>
because to access the files I would need to unify the user/group of both services (if DynamicUser even works across services, never used it before this)
endformationage has joined #nixos
<ottidmes>
and then workaround actually wanting to probably use a different user/group when storing the backup
<symphorien>
sorry for leading you to this trap
JosW has quit [Ping timeout: 252 seconds]
<ottidmes>
not at all, learned something new and I fully agreed with the reason behind using it, but I will revert it back to reserving an uid/gid
thedavidmeister has joined #nixos
<thedavidmeister>
hi
<thedavidmeister>
how do i fix a corrupt manifest?
<thedavidmeister>
/nix/store/f4g5yfc1yjm7lhi03x93zm8crl3ckr7b-env-manifest.nix is empty
<ottidmes>
thedavidmeister: you could also try and see if you can remove all roots to it (deinstalling it from nix-env/nixos, removing relevant result symlinks), and then nix-store --delete it, and rebuild it
<thedavidmeister>
how do i remove it?
<ottidmes>
I just told you, first remove its roots, and then you can remove it with nix-store --delete
<tilpner>
thedavidmeister: But whatever you do, do not ignore liveness
<thedavidmeister>
"remove its roots"
<thedavidmeister>
i don't know what that means
<tilpner>
Good
<ottidmes>
"deinstalling it from nix-env/nixos, removing relevant result symlinks"
<thedavidmeister>
what is "it"?
<thedavidmeister>
this started happening after i ran 'nixos-rebuild switch --upgrade' and my laptop ran out of power
yayforj has joined #nixos
<ottidmes>
whatever your env-manifest.nix is
<ottidmes>
thedavidmeister: nix-store --query --roots /nix/store/f4g5yfc1yjm7lhi03x93zm8crl3ckr7b-env-manifest.nix, what does that give you
<thedavidmeister>
how do i know what my env-manifest.nix is?
<thedavidmeister>
ill try
<ottidmes>
thedavidmeister: its apparantly something linked to a profile, I checked my nix store, and I have one for my system profile and user profile
<ottidmes>
I assume yours is from your system profile, considering your last command
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ottidmes>
you are right, just checked the manpage, and nix-store --verify does not expect parameters
knupfer has quit [Remote host closed the connection]
<ottidmes>
without parameters it did not work, right? then the only thing I can think of is to call nixos-rebuild switch --upgrade again and hope you get a new channels-4-link
<ottidmes>
nix-channel --upgrade is the same I think (as --upgrade for nixos-rebuild)
rcshm has quit [Remote host closed the connection]
<thedavidmeister>
error: program '/nix/store/fz9lwpq4ikv8m6spwzcw030kdgjh0sm4-nix-2.1.3/bin/nix-env' failed with exit code 1
<thedavidmeister>
what do you mean wait with the updating?
<ottidmes>
well since repair doesnt work, my idea was to rollback to a place where things were not corrupt, and everything new after that could than hopefully be GC'ed, thinking that if you were to update again, it would not try and rebuild things already in the store (including the corrupt paths), our goals is to erradicate those corrupt paths, and since repair doesnt work, rollback and then GC was my next idea
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<thedavidmeister>
oh
<thedavidmeister>
is there a way to delete the paths?
<ottidmes>
yes, but only if there are no live roots to it
<thedavidmeister>
ah, what is the command to find live roots?
<ottidmes>
like I said before, you can delete paths safely with nix-store --delete, but that will fail with live roots to the path (rightfully so)
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<ottidmes>
check the nix-store manpage under nix-store --gc
<ottidmes>
nix-store --gc --print-live should show you all live roots
petrkr has quit [Remote host closed the connection]
<ottidmes>
nix-store --query --roots /nix/store/path... shows the roots (live or not) of any /nix/store path
<thedavidmeister>
ok
<thedavidmeister>
still says channels-3-link
Glider_IRC__ has quit [Ping timeout: 246 seconds]
daniele- has quit [Quit: daniele-]
<ottidmes>
because you did nix-channel --update
maximiliantagher has quit [Ping timeout: 240 seconds]
<{^_^}>
nix#1281 (by ocharles, 1 year ago, open): nix-store --repair-path isn't repairing, manually repairing works
o1lo01ol1o has joined #nixos
<ottidmes>
and maybe try nix-store --delete /nix/store/25xwf42prc2ynhgyv1p6qh5dcf6fjzcz-nixos-19.03pre168613.ffc604e5579.drv, maybe that way it will be regenerated
<thedavidmeister>
can't because it is still alive
<thedavidmeister>
i saw some ppl saying delete the profile manually
daniele- has joined #nixos
shibboleth has quit [Quit: shibboleth]
Rusty1 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<ottidmes>
you have to be very careful with that, you can make things a lot worse that way
<ottidmes>
now you have corrupt files, but by doing that you might get a corrupt database
freeman42y has quit [Ping timeout: 252 seconds]
sicklork1n has joined #nixos
<thedavidmeister>
ok i just moved it
<{^_^}>
[nixpkgs] @rnhmjoj opened pull request #55495 → nix-script: init at 2015-09-22 → https://git.io/fhHHO
freeman42x has joined #nixos
maximiliantagher has joined #nixos
eadwu has quit [Quit: WeeChat 2.3]
sicklork1n has quit [Read error: Connection reset by peer]
rfold has quit [Quit: leaving]
drakonis has quit [Quit: WeeChat 2.3]
maximiliantagher has quit [Ping timeout: 268 seconds]
thc202 has quit [Ping timeout: 246 seconds]
<thedavidmeister>
kk seems to be working better...
<thedavidmeister>
thanks heaps
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
htoc^ has quit []
<dunk>
How can I force an install to my homedir, and not elsewhere on the system?
<lewo>
dunk: if you nix-env -i with your user, the package is only avalaible for your user (but files are created in /nix/store)
ottidmes has quit [Ping timeout: 244 seconds]
<dunk>
lewo: on my system there seems to be the following symlink: ~/.nix-profile -> /nix/var/nix/profiles/default
drakonis has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]