<evils>
looks like wrapGAppsHook is what i needed, and buildInputs
toppler has joined #nixos
<evanjs>
Well this is odd. Disabled and re-enabled openrazer-daemon and now the ghost service is gone, but systemd still can't "find" the service file lol
JonReed has quit [Ping timeout: 260 seconds]
<evanjs>
Checked my work computer and it looks like it's running fine over there >.>
<CMCDragonkai>
I noticed that when building nixos images or netboot. The version is always `pre130979.gfedcba`. This is apparently hardcoded into the repo at nixos/release.nix as `revCount` and `shortRev` parameters. Are these parameters meant to be set by the person running the build?
mdash has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 pushed 3 commits to bash-no-undef-vars: https://git.io/Jeat0
<clever>
CMCDragonkai: they are usually set by hydra, when it imports release.nix
soju has joined #nixos
gagbo has quit [Ping timeout: 276 seconds]
<CMCDragonkai>
clever: can you show me which script sets it? I would like to set up something similar so I get the proper revcount and hash.
gagbo has joined #nixos
m0rphism has quit [Ping timeout: 265 seconds]
<CMCDragonkai>
Right now the built netboot image has the wrong revcount/hash, revcount not so important, but at least with hash, that's definitely not the nixpkgs that is bundled into the netbooted image
eacameron has quit [Quit: Connection closed for inactivity]
waleee-cl has quit [Quit: Connection closed for inactivity]
<evanjs>
Yeah this openrazer thing is definitely something with one computer. Works on both my other NixOS machines... what even
jb55 has joined #nixos
jluttine has quit [Ping timeout: 240 seconds]
lovesegfault has joined #nixos
toppler has joined #nixos
ddellacosta has quit [Ping timeout: 240 seconds]
mexisme has joined #nixos
jgt has joined #nixos
jgt has quit [Ping timeout: 246 seconds]
drakonis has joined #nixos
lovesegfault has quit [Quit: WeeChat 2.6]
hlolli__ has joined #nixos
lovesegfault has joined #nixos
lovesegfault has quit [Client Quit]
hlolli_ has quit [Ping timeout: 268 seconds]
lovesegfault has joined #nixos
silver has quit [Read error: Connection reset by peer]
lovesegfault has quit [Client Quit]
lovesegfault has joined #nixos
bvdw has quit [Quit: bvdw]
bvdw has joined #nixos
leungbk has joined #nixos
<infinisil>
Watching a nixos-rebuild is really deceiving
<infinisil>
It makes you feel like you're doing somewhere, even though you're just staring at it
<infinisil>
s/somewhere/something
<infinisil>
s/nixos-rebuild/long nix-build
<leungbk>
How can I make an environment with gcc in which enableMultilib is true? I've tried googling but couldn't figure out how to write the shell.nix properly.
ertz has joined #nixos
bvdw has quit [Quit: bvdw]
<evanjs>
infinisil: nix-top can be slightly better *shrugs*
growpotkin has joined #nixos
bvdw has joined #nixos
<evanjs>
>> when a systemd service works fine on 2/3 NixOS machines, though >.> time to increase systemd output verbosity...
Rusty1 has joined #nixos
Guest84 has joined #nixos
<CMCDragonkai>
clever: I tried rebuilding while specifying the parameters `nix-build nixos/release.nix --attr netboot.x86_64-linux --arg nixpkgs "{ outPath = ./.; revCount = $revCount; shortRev = \"$shortRev\"; }"`
<CMCDragonkai>
I'd like to use `cleanSource` as it is used in `release.nix`
leungbk has quit [Quit: ERC (IRC client for Emacs 26.3)]
andromeda-galaxy has joined #nixos
ddellacosta has joined #nixos
toppler has quit [Ping timeout: 240 seconds]
smw_ has joined #nixos
jluttine has joined #nixos
jgt has joined #nixos
<{^_^}>
[nixpkgs] @redvers opened pull request #72869 → icu: bugfix #71142 - Fixes icu-config reporting bad location → https://git.io/Jeaq9
<smw_>
I want to setup strongswan with custom config. Specifically, I want to change charon.kernel-netlink.fwmark. To do this in nixos, will I need to completely rewrite the strongswan module?
<smw_>
Interestingly, strongswan-swanctl has strongswan.extraConfig but is much harder to use. For example, I haven’t figured out what the equivalent of ike= in a IPSec.conf conn.
evanjs has joined #nixos
jgt has quit [Ping timeout: 252 seconds]
<__red__>
So - I have a new package I want to add to nixpkgs but it's not going to build until the (PR) bugfix I just opened is committed. I'm guessing that I should probably wait until my bugfix PR is applied before pushing the new package up for a PR?
<__red__>
(it's ltfs - LTO filesystems) :-)
<evanjs>
tilpner_: lol I think you won. Traced it back to ~/.config/systemd/user and hey look, a symlink to openrazer-daemon from september 7th
<evanjs>
everything else is from today...
<evanjs>
Oh hey look, I removed that file and all works fine now >.> if only I was told about ~/.config/systemd/user r in the first error message lol
toppler has joined #nixos
<{^_^}>
[nixpkgs] @nh2 opened pull request #72870 → WIP: Add RNNoise, use it in mumble → https://git.io/Jeaqb
hlolli__ has quit [Remote host closed the connection]
inkbottle has quit [Ping timeout: 268 seconds]
hlolli__ has joined #nixos
inkbottle has joined #nixos
<{^_^}>
[nixpkgs] @aanderse opened pull request #72871 → zabbix: minor version bump → https://git.io/JeaqA
o1lo01ol1o has joined #nixos
h0m1 has quit [Ping timeout: 276 seconds]
mexisme has quit [Ping timeout: 240 seconds]
h0m1 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
vika_nezrimaya has quit [Ping timeout: 265 seconds]
pareidolia has joined #nixos
<Ankhers>
Does anyone have experience using unity3d on nixos? Specifically I am hoping to use unity hub. It is basically a program that downloads and manages multiple versions of unity and your different projects.
Guest84 has quit [Ping timeout: 265 seconds]
dreverri has quit [Remote host closed the connection]
<Ankhers>
Actually, more generally, how does nixos deal with applications that want to update themselves?
<{^_^}>
[nixpkgs] @marsam opened pull request #72872 → emacs: also allow librsvg dependency when using withNS → https://git.io/Jeam7
wildtrees has quit [Quit: Leaving]
<mdash>
Ankhers: your choices are "don't keep them in the store" or "generate nix exprs for each update" or some combination thereof
<Ankhers>
mdash: Don't keep them in the store, as in don't install them with nix? So just act like it was any other linux distro for those applications?
<notgne2>
has anybody had any luck building pcsx2 on 19.09
lovesegfault has joined #nixos
lovesegfault has quit [Client Quit]
lovesegfault has joined #nixos
mexisme has joined #nixos
lovesegfault has quit [Client Quit]
lovesegfault has joined #nixos
gagbo has quit [Ping timeout: 240 seconds]
toppler has joined #nixos
gagbo has joined #nixos
<wedens[m]>
zeta_0: yes. just add python to `home.packages` instead of `environment.systemPackages`
<lovesegfault>
I'm trying to run the update script for git in nixpkgs, but I get this: error: anonymous function at /home/bemeurer/src/nixpkgs/pkgs/applications/version-management/git-and-tools/git/default.nix:1:1 called without required argument 'fetchurl', at (string):1:6
<lovesegfault>
Can someone explain me what's going on?
drakonis has joined #nixos
<zeta_0>
wedens[m]: what is the syntax to do something similar, like with haskell ?
<zeta_0>
wedens[m]: (haskell.packages.ghc865.ghcWithHoogle (hpkgs: with hpkgs; [
<zeta_0>
wedens[m]: thanks for the sending me the link
spacefrogg has joined #nixos
smw_ has quit [Quit: Computer has gone to sleep]
smw_ has joined #nixos
growpotk- has joined #nixos
xkapastel has joined #nixos
lovesegfault has quit [Quit: WeeChat 2.6]
lovesegfault has joined #nixos
mexisme has quit [Ping timeout: 268 seconds]
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
justanotheruser has quit [Ping timeout: 265 seconds]
<ivan>
can I have multiple extension dirs in vscode/vscodium where one is writable? nixpkgs is setting a read-only extension dir when I enable any extensions
<cjpbirkbeck>
the manual should probably be clearer on the point
<DigitalKiwi>
had to find it in my log and then had to find it in the online log and had the date wrong ;_;
<samueldr>
the part I quoted in that logs link is from the manual
<samueldr>
>> Since ${ and '' have special meaning in indented strings, you need a way to quote them. $ can be escaped by prefixing it with '' (that is, two single quotes), i.e., ''$. '' can be escaped by prefixing it with ', i.e., '''. $ removes any special meaning from the following $. Linefeed, carriage-return and tab characters can be written as ''\n, ''\r, ''\t, and ''\ escapes any other character.
<samueldr>
though yeah, it could get some improvement, mainly shown in a table
<samueldr>
columns of "strings, indented strings, desired result" and rows for each desired result, with an empty cell when it doesn't need to be escaped
Acou_Bass has quit [Ping timeout: 265 seconds]
<DigitalKiwi>
nobody has time to figure out how to escape that many levels
<infinisil>
Ideally editors would escape everything you type by default (depending on context)
<infinisil>
And have some shortcut you can use to have things not escaped
<samueldr>
Ideally programming shouldn't be done in a character soup
<infinisil>
Hehe yes
<infinisil>
AST editors ftw
Acou_Bass has joined #nixos
dansho has quit [Ping timeout: 265 seconds]
<infinisil>
But I like how zsh escapes things I type automatically
<DigitalKiwi>
as soon as they figure out whether tab key means 2 spaces 3 spaces 4 spaces a literal \t or some arbitrary amount based on context...
<infinisil>
Actually not sure what causes it, but when I type "http://" then ";" in my terminal it becomes "http://\;"
<infinisil>
Only happens in urls though
<Miyu-saki>
> AST editors
<{^_^}>
undefined variable 'AST' at (string):271:1
<Miyu-saki>
I see this is how Lisp starts to live again.
<DigitalKiwi>
coda and unison are worth looking at
<infinisil>
I never tried out paredit, but I heard it's pretty good
domogled has joined #nixos
<wedens[m]>
unison is not ast editor though
<DigitalKiwi>
isn't it ast source code? tbh it's above my head
<Miyu-saki>
infinisil: I've used Paedit. It's amazing for Nix.
<Miyu-saki>
Paredit*
dansho has joined #nixos
<wedens[m]>
no. on the outside it's just a regular programming language. I believe there were some plans to make a visual editor for unison, but Paul have decided to do without it
endformationage has quit [Ping timeout: 268 seconds]
toppler has joined #nixos
ddellacosta has quit [Ping timeout: 265 seconds]
Rusty1 has quit [Quit: Konversation terminated!]
drakonis has quit [Ping timeout: 240 seconds]
drakonis has joined #nixos
orivej has joined #nixos
MarcelineVQ has joined #nixos
MarcelineVQ has left #nixos [#nixos]
zupo has joined #nixos
Okinan has quit [Quit: Konversation terminated!]
kenran has joined #nixos
jedai42 has joined #nixos
lovesegfault has quit [Ping timeout: 246 seconds]
cjpbirkbeck has quit [Quit: Quitting now.]
lovesegfault has joined #nixos
orivej has quit [Ping timeout: 265 seconds]
bahamas has joined #nixos
domogled has quit [Ping timeout: 265 seconds]
b1000101 has quit [Read error: Connection reset by peer]
b101010 has joined #nixos
<bahamas>
when creating a nix package from a Python package, does nix always put the Python package inside a Python interpreter? is there a way to avoid that?
mexisme has joined #nixos
toppler has quit [Ping timeout: 268 seconds]
orivej has joined #nixos
<pie_>
is skypeforlinux broken or is it just me
<pie_>
im guessing the binaries arent in the cache
orivej has quit [Ping timeout: 268 seconds]
toppler has joined #nixos
xd1le has joined #nixos
palo1 has joined #nixos
palo has quit [Ping timeout: 240 seconds]
palo1 is now known as palo
Kolby__ has quit [Read error: Connection reset by peer]
Kolby__ has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos
zupo has quit [Client Quit]
Acou_Bass has quit [Ping timeout: 265 seconds]
b1000101 has joined #nixos
lovesegfault has quit [Ping timeout: 264 seconds]
bahamas has quit [Ping timeout: 265 seconds]
lovesegfault has joined #nixos
b101010 has quit [Ping timeout: 240 seconds]
hisham has quit [Ping timeout: 240 seconds]
rauno has joined #nixos
tmaekawa has joined #nixos
CMCDragonkai has quit [Quit: Connection closed for inactivity]
<DigitalKiwi>
easy there dimmesdale it's not always obvious where to find the documenation for everything being as it's often incomplete out of date or spread out over multiple manuals :P
<wedens[m]>
Ariakenom: no idea. if you feel like investigating, you can clone the library and play with it :)
<DigitalKiwi>
is this a question i find the answer in the nix manual, nixos manual, nixpkgs manual, nixos.wiki, nixos options, nixos packages, random nix* --help/man, obscure blog, reading source, trial and error, ...
<Ariakenom>
I will just sympathise with Synthetica9 and their struggle for now
<Synthetica>
Ariakenom: You could always use sudoedit
<bahamas>
I see that when packaging a Python package, nix puts it inside a python interpreter's site-packages. does nix offer a way to refer to that package's path? I want to avoid doing "${package}/python3.7/site-packages/my-package/static_all"
<Ariakenom>
I have been using emacs. It would just be slightly smoother if this worked
<wedens[m]>
bahamas: what are you trying to do? package set for specific interpreter can be referenced via `python37Packages`, for example
<Ariakenom>
what does sudoedit do?
orivej has quit [Ping timeout: 265 seconds]
<wedens[m]>
bahamas: also `python37.pkgs`
<wedens[m]>
if that's what you're asking
vonfry has quit [Quit: WeeChat 2.6]
<bahamas>
wedens[m]: I'm packaging my project with nix and want to deploy it to NixOS. also, I want to serve the static files from my project with nginx. in order to do that, I need to point to mypackage/static_all directory
<wedens[m]>
you may want to create a separate output for static files. or separate derivation
toppler has quit [Ping timeout: 240 seconds]
<bahamas>
wedens[m]: do you have any pointers on how to do that?
zupo has joined #nixos
<wedens[m]>
bahamas: do you use `buildPythonPackage` for your package?
alexherbo2 has joined #nixos
<wedens[m]>
anyway, you can add additional output: `outputs = [ "out" "static" ];`
<wedens[m]>
it can be referenced the same way, you reference 'out': `$static`
<bahamas>
wedens[m]: yes, I'm using buildPythonPackage
<wedens[m]>
then you can copy directory to this output in some phase. for example, `postFixup = "cp -r ./static $static";`
<wedens[m]>
and it can be referenced as `python37Packages.my-package.static`
LysergicDreams has quit [Ping timeout: 240 seconds]
<wedens[m]>
as I said earlier, alternative is to make a new derivation with static.
<bahamas>
wedens[m]: you mean to create a package.nix file just for the static dir and pass that in propagatedBuildInputs?
<wedens[m]>
why do you think you need `propagatedBuildInputs`?
LysergicDreams has joined #nixos
<tetdim>
pwl
<bahamas>
wedens[m]: you're right. I don't. but I need to reference this derivation in my package somehow, so I can use it in configuration.nix
<wedens[m]>
I don't think I understand what you intend to do with the static
toppler has joined #nixos
<gchristensen>
bahamas: it sounds like sharing a code sample might help
<wedens[m]>
if you want to serve it via nginx or something, as I said earlier, you can reference it via `myStaticFiles = "${python37Packages.my-package.static}"`
LysergicDreams has quit [Ping timeout: 250 seconds]
<bahamas>
wedens[m]: gchristensen I'm doing this in my configuration.nix
<bahamas>
wedens[m]: I'm assuming from how you reference python37Packages.my-package, that I should add my package with an overlay?
LysergicDreams has joined #nixos
<bahamas>
currently, I have a service.nix where I define the service for my package. I add this service.nix to `imports` in configuration.nix
<Synthetica>
Ariakenom: It creates a temporary copy of a file, opens it with EDITOR, waits until done, and copies it back
<Synthetica>
So your editor never has full sudo rights
Zer0xp has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
<Synthetica>
(You should have EDITOR=code --wait, because otherwise it forks and returns immediately, which is not what you want)
mounty has joined #nixos
tilpner_ is now known as tilpner
Zer0xp has quit [Quit: Bye!]
<bahamas>
wedens[m]: your suggestion with adding a 'static' output worked. thank you!
asymmetric has quit [Read error: Connection reset by peer]
vuko has quit [Quit: leaving]
vuko has joined #nixos
philr has quit [Ping timeout: 240 seconds]
Zer0xp has joined #nixos
ThatDocsLady has joined #nixos
<mrSpec>
Hi! What is the correct way of removing declaratively defined nixos container? I've removed it from configuration, switched, did gc -d. I've created new nixOS container with the same name... and I can still see my old commands in bash history o_O
ng0 has joined #nixos
<tilpner>
mrSpec: It might keep around the container data, to prevent loss of important data
<tilpner>
mrSpec: And re-defining the container just uses that data again
<mrSpec>
tilpner: I've tried destroy but got /run/current-system/sw/bin/nixos-container: cannot destroy declarative container (remove it from your configuration.nix instead)
<mrSpec>
so should I manualy remove the dir?
<tilpner>
Yes, try that
cosimone has joined #nixos
<mrSpec>
thanks, done.
jgt has quit [Ping timeout: 246 seconds]
mounty has quit [Quit: Konversation terminated!]
mounty has joined #nixos
Zer0xp has quit [Quit: Bye!]
toppler has joined #nixos
mounty has quit [Client Quit]
mounty has joined #nixos
iyzsong has joined #nixos
xkapastel has quit [Quit: Connection closed for inactivity]
vika_nezrimaya has joined #nixos
ThatDocsLady_ is now known as ThatDocsLady
<pie_>
> pkgs.(hello)
<{^_^}>
error: syntax error, unexpected '(', expecting ID or OR_KW or DOLLAR_CURLY or '"', at (string):271:6
<{^_^}>
[nixpkgs] @cdepillabout opened pull request #72882 → haskellPackages.purescript: get this building again → https://git.io/JeanE
bahamas has quit [Ping timeout: 245 seconds]
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
ilikeheaps has quit [Ping timeout: 252 seconds]
kenran has quit [Ping timeout: 265 seconds]
<Twey>
How does `stdenv` get into `pkgs`? I see where the stdenv is built (pkgs/stdenv/{generic,linux,…}/default.nix) and I see `pkgs.stdenv` being referenced in `pkgs/top-level/all-packages.nix`, but I don't see how it gets into the package set in the first place
<tilpner>
Twey: Check stage.nix
<{^_^}>
[nix] @edolstra pushed to flakes « Use revcount/last-modified for computing the flake fingerprint »: https://git.io/JeanK
sigmundv has joined #nixos
<eyJhb>
tilpner: thanks :D
jybs has joined #nixos
fendor has joined #nixos
<etu>
tilpner: oh, that's how I get rid of that :D
is_null has quit [Ping timeout: 268 seconds]
<Twey>
tilpner: Oho, thank you!
Franciman has joined #nixos
<Franciman>
Hi
<{^_^}>
[nixpkgs] @adisbladis merged pull request #72872 → emacs: also allow librsvg dependency when using withNS → https://git.io/Jeam7
<xwvvvvwx>
Franciman: If you prefer to use scheme, then maybe just use Guix, the project is quite nice ;)
Chiliparrot has joined #nixos
<Franciman>
it's still a bit young, tho
cosimone has joined #nixos
<Franciman>
I do love guix, but I needed something right now for production
<Franciman>
nix is so messy :<
<Franciman>
if there is no way tho, no problem
<Franciman>
thanks
<infinisil>
Franciman: What's your use case? What are you considering using dhall for?
sb0 has quit [Quit: Leaving]
bahamas has joined #nixos
<{^_^}>
[nixpkgs] @fooker opened pull request #72883 → nixos/networkd: add `Name` as valid option to `Link` → https://git.io/Jeacs
<{^_^}>
[nixpkgs] @fooker closed pull request #72863 → nixos/networkd: add `Name` as valid option to `Link` → https://git.io/JeaIi
toppler has quit [Ping timeout: 268 seconds]
<Franciman>
infinisil, writing configurations
<Franciman>
I don't want to use nix's language
<{^_^}>
[nixpkgs] @fooker opened pull request #72884 → nixos/networkd: Fix allowed values for IPv6PrefixDelegation → https://git.io/Jeacc
<Franciman>
i saw that dhall has a backend for nix
<kirelagin>
dhall comes with a dhall-to-nix compiler, you can just use it, if you want
fendor has joined #nixos
<Franciman>
yes, I wanted to know whether somebody found it usable
vika_nezrimaya has quit [Ping timeout: 265 seconds]
<Franciman>
or used it with profit
<kirelagin>
dhall? or what?
<Franciman>
dhall
<Shoubit>
Franciman, IIRC you can use Dhall for some things but I believe Dhall's type-system doesn't support all of what Nix does, and so you may have some issues.
<kirelagin>
The compiler works, therefore, I guess, it qualifies as “usable”
<Franciman>
Shoubit, oh i see
<Franciman>
yeah that was things I was interested into
<Franciman>
so kirelagin not really super usable
<Franciman>
ok thanks ppl
<kirelagin>
Usable for what? You said, you wanted configuration, dhall is usable for configuratoin, as it is a configuratoin langauge.
<gchristensen>
sounds like Franciman wants to use dhall to emit a configuration.nix
<Franciman>
exactly
<kirelagin>
Nix, OTOH, is a programming language, which dhall is not. So you can use dhall to write your configuration and then use the dhall-to-nix- compile, to compile your dhall configuration to Nix configuration.
<kirelagin>
dhall should be perfectly enough for `configuration.nix` as long as you don’t want any logic (that is, functions) in it.
<gchristensen>
which is perfectly good enough for as long as it is good enough :)
<Franciman>
hmm
<Franciman>
ok i understand
<wedens[m]>
nix is programming language to the same degree dhall is
<kirelagin>
well, I mean, if you don’t like Nix and are looking for an alternative, there has to be a price for you to pay
<Franciman>
it's better to continue with the ugly nix
<gchristensen>
Franciman: I would agree with that, as we'll be able to better help
<Franciman>
you're right kirelagin
<Franciman>
I guess at this stage the lower price I can pay is the ugliness of nix
<Franciman>
the other prices are far too high
<Twey>
kirelagin: Dhall isn't a programming language?
<Franciman>
it's not turing complete
<Franciman>
so you can't loop forever
<wedens[m]>
both dhall and nix are configuration languages
<kirelagin>
The README says it is a configuration language, so, not, it is not a programming language
<wedens[m]>
not a general puropse languages
<Franciman>
i think you can do computations that terminate
<Franciman>
(not all, a subclass)
<qyliss>
You can.
<kirelagin>
Well, the obvious difference is that dhall is typed, so what you can do in dhall is a subset of what you can do in nix. And what you can do in dhall, that is, its design, is informed by the idea that it is a “better YAML”, not a programming language.
<Twey>
Other programming languages that aren't Turing-complete: Agda, C :þ I buy ‘not a general-purpose language’ (like Nix, it clearly has an intended special purpose) but ‘not a programming language’ is a pretty hard claim to back up for most things
<Twey>
But eh, words
<kirelagin>
Twey: C?
<kirelagin>
Agda is not a programming language.
<qyliss>
"what you can do in dhall is a subset of what you can do in nix." doesn't follow from "dhall is typed"
<Twey>
>Agda is a dependently typed programming language / interactive theorem prover.
<kirelagin>
it does
<wedens[m]>
"Agda is not a programming language." lol?
<Franciman>
anyways I didn't mean to start this off topic
<Franciman>
sorry
<qyliss>
Rust and Haskell are typed, but no less powerful than JavaScript or Perl
<Twey>
kirelagin: There are fiddly bits of the C spec that make it not Turing-complete (there are necessarily a finite number of possible states)
<gchristensen>
let's take this chat to -chat :P
<Twey>
Yeah, sorry :)
<kirelagin>
Well, I mean, Agda is a proof assistant first and foremost. I don't know about these days, but if you tried to write an actual “program” (that is, something that executes, performs a computation, and outputs a result), you had to SUFFER. For me that doesn't sound like a general-purpose programming language.
<kirelagin>
Right, sorry :)
lortabac has joined #nixos
f-a has joined #nixos
<Twey>
If I'm defining an overlay, and I refer to a package in `super`, does that package already have its dependencies overridden with the packages from my overlay?
f-a has left #nixos [#nixos]
bahamas has quit [Ping timeout: 240 seconds]
toppler has joined #nixos
kenran has joined #nixos
Soo_Slow has joined #nixos
lortabac has left #nixos ["WeeChat 1.9.1"]
rauno has quit [Ping timeout: 240 seconds]
<wucke13>
Why do so many options for password files refer /run/keys/*, when that defaults to a ramfs
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<infinisil>
wucke13: You mean the example values?
jybs has quit [Read error: Connection reset by peer]
<wucke13>
Yes
rauno has joined #nixos
captn3m0- has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @alyssais pushed 3 commits to release-19.09: https://git.io/Jeacp
<infinisil>
These files would be put there with nixops or so, and a ramfs makes sure they're not written to disk
<Twey>
Why does the former work while the latter results in an error?
<wucke13>
What would be a good choice for something which is not deployed via nixops?
FRidh has joined #nixos
<infinisil>
Twey: define "does not work"
<Twey>
I was trying to demonstrate with the bot but I think I failed… let me pastebin it
<Twey>
https://hastebin.com/simevebepa.nix — the specific behaviour isn't that important… essentially, I expected these two things to be the same, but it seems like the value of ‘super’ is very different
justanotheruser has quit [Ping timeout: 268 seconds]
<infinisil>
Twey: Put a space between the ">" and the expression for the bot
captn3m0 has joined #nixos
<infinisil>
Twey: Ah, I think this doesn't work because the overlay is used by nixpkgs itself, and there's lots of things that need stdenv
<Twey>
infinisil: So those things are not equivalent? `overlays` extends the package set used internally by nixpkgs, while `extend` only applies ‘after the fact’ in some sense?
<infinisil>
Hmm...
<{^_^}>
[nixpkgs] @alyssais pushed commit from @r-ryantm to release-19.03 « zsh-history-substring-search: 1.0.1 -> 1.0.2 »: https://git.io/JeaC3
<{^_^}>
[nixpkgs] @JohnAZoidberg opened pull request #72886 → manim: init at 0.1.10 → https://git.io/JeaCG
<infinisil>
Yeah now that you say that I'm wondering too..
abathur has quit [Ping timeout: 240 seconds]
Ariakenom has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @domenkozar merged pull request #72833 → Port tests to Python → https://git.io/Je2Nz
<{^_^}>
[nixpkgs] @1000101 opened pull request #72887 → Port tests to Python → https://git.io/JeaC4
<{^_^}>
[nixpkgs] @fooker opened pull request #72888 → nixos/networkd: Add more valid values for RequiredForOnline → https://git.io/JeaCB
Soo_Slow has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @curiousleo opened pull request #72889 → lorri (including user service): init at version rolling-release-2019-10-30 → https://git.io/JeaCa
Franciman has left #nixos ["Leaving"]
bahamas has joined #nixos
cosimone has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @Ma27 pushed commit from @abbradar to release-19.09 « matrix-synapse service: blacklist local IPv6 addresses by default »: https://git.io/JeaCM
arjen-jonathan has joined #nixos
zupo has joined #nixos
<niksnut>
strange, the nlohmann_json tests should only take a few minutes
cosimone has quit [Quit: Terminated!]
<gchristensen>
niksnut: several people on that project report it taking a long time
<niksnut>
maybe we should just disable the tests?
<gchristensen>
yeah that might do
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<davidak>
at least for now
<davidak>
until the upstream issue is resolved
silver has joined #nixos
<exarkun>
What about reverting the package upgrade?
<niksnut>
what upgrade?
<exarkun>
The nlohmann_json upgrade?
<niksnut>
I don't think there is an upgrade on the stable channel
toppler has joined #nixos
<niksnut>
it has been at 3.6.1 since march
<exarkun>
Upgrade of one of its dependencies?
<niksnut>
it's a timeout, so it's probably a random thing
<exarkun>
Is the timeout non-deterministic and it just coincidentally never happened until 5 days ago and then consistently happened for the last 5 days?
<niksnut>
which is why we shouldn't run package tests btw
<exarkun>
"don't run the tests" is a rather broad conclusion, I think
<niksnut>
well, the odds of a channel rebuild succeeding converges to 0 the more upstream tests we run
<gchristensen>
michaelraskin has a good idea on this, instead of do-run-tests and dont-run-tests, make it a tristate of known-to-work, known-to-be-broken, unkown
kleschenko has joined #nixos
<xd1le>
are nix flakes, "require.nix"?
monotux has quit [Quit: WeeChat 2.6]
monotux has joined #nixos
<niksnut>
oh, the failure is on aarch64
<niksnut>
I guess the tests don't take a few minutes there
averell has quit [Remote host closed the connection]
<gchristensen>
I wonder why not
<gchristensen>
does it need more cores?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<exarkun>
boot-once will try the new configuration once, it will fail, and when the system is rebooted it will go back to the previous configuration which worked and the system will boot
<exarkun>
yea?
<tilpner>
Oh, this is a headless case, where you can't select the previous generation?
<exarkun>
I'm thinking about remote server management, yea, which is often kind of like headless
<exarkun>
A service ticket that says "reboot server 1234" might get handled in 5 minutes, a service ticket that says something more complicated might take a few hours or a day
<exarkun>
I mean, ideally, something would notice that the system has wedged and is dead and *automatically* do the reboot back to working configuration :) but I'm not aware of any such functionality
<gchristensen>
you can use systemd watchdog to issue reboots if things are wedged, and you can specify default boot entries
ddellacosta has joined #nixos
<exarkun>
ooh nice
<gchristensen>
iirc lumiguide does this, and another co deploying nixos to remote corners of earth
<exarkun>
so yea now I want `nixos-rebuild switch-failsafe` to set up a hardware watchdog to reboot back into the current configuration if the new one fails and make the new one permanent if it doesn't
<gchristensen>
me too
elais[m] is now known as elais[m]1
<gchristensen>
sometimes I setup a scheduled reboot with shutdown -r if I have a risky one, and use nixos-rebuild test
dreverri has quit [Remote host closed the connection]
mexisme has quit [Ping timeout: 264 seconds]
elais[m]1 is now known as elais[m]
<exarkun>
yea, I did that a couple times yesterday
dreverri has joined #nixos
is_null has joined #nixos
<dminuoso>
Mmm. So Im in the situation where Im essentially forced to keep secrets in the nix store unless I want to rewrite some modules. In principle this is not a shared machine, so people that have access to it have access to any secrets anyway.
<dminuoso>
Is it reasonable enough to just not care and keep secrets in the nix store?
<gchristensen>
dminuoso: I do that sometimes yes, but the most professional response I could give is don't put secrets in the store
<rnhmjoj>
are there any haskell and nix developers here? i'm quite puzzled on the state of things regarding cabal, nix integration and v2/new style flags
is_null has quit [Remote host closed the connection]
elais[m] has left #nixos ["Kicked by @appservice-irc:matrix.org : issued !quit command"]
<exarkun>
dminuoso: You should write down your threat model
<dminuoso>
rnhmjoj: Yes
<dminuoso>
rnhmjoj: My gut feeling says that likely a third in here are Haskell developers based on the names I regularly see talking...
<exarkun>
dminuoso: Include it in your vulnerabilities (like secrets in /nix/store) and your defenses (like controlled access to the system)
<dminuoso>
exarkun: Ah thats not a bad idea.
<dminuoso>
rnhmjoj: In short, v2- is incompatible with nix (although you can use nix to set up cabal, and then just do v2-*)
<dminuoso>
rnhmjoj: You'd use cabal v1- orchestrated by nix, with cabal2nix you can generate the appropriate nix derivations for your packages easily.
dreverri has quit [Ping timeout: 268 seconds]
<rnhmjoj>
dminuoso: thank you. so.. does v2 still works using the good old cabal2nix workflow?
<dminuoso>
rnhmjoj: No. v2- is basically an imitation of what nix does, but with a scope on only the haskell side of things.
<rnhmjoj>
dminuoso: ah, that's bad. does that mean when v1- will be deprecated i'll have to move to this new method?
<dminuoso>
rnhmjoj: I dont think v1- will be deprecated. :)
<dminuoso>
rnhmjoj: It's just that at some point `cabal build` will, by default, invoke `v2-build`
<rnhmjoj>
dminuoso: so, right now cabal build is still the v1-build?
<dminuoso>
rnhmjoj: Perhaps? Im not sure actually :P
<infinisil>
Cabal 3.0 switched to new-syle builds by default
<infinisil>
Which is already released
<infinisil>
> cabal-install.version
<{^_^}>
"3.0.0.0"
<exarkun>
is there NixOS configuration for the systemd udevd mdadm stuff that seems to be set up (by default, I guess)? I'm not using md at all and I'm kind of uneasy seeing systemd-udevd running md commands on my zfs array disks and spewing errors.
captn3m0 has joined #nixos
<dminuoso>
infinisil: Ah! I use cabal 3.0 - but I always issue v1- and v2- prefixes explicitly, so I really dont know :)
abathur has quit [Ping timeout: 240 seconds]
<dminuoso>
Just know that this move was planned.
<rnhmjoj>
infinisil: uhm, i'm using 3.0.0 but cabal commands (unprefixed) seem to work as usual inside a nix-shell
ajirx has quit [Quit: Leaving]
<infinisil>
It's cabal the executable (from cabal-install) that dies this
<infinisil>
does*
<dminuoso>
I wonder whether thats perhabs configurable in ~/.cabal/config
Neo-- has quit [Ping timeout: 240 seconds]
<dminuoso>
rnhmjoj: What do you mean by "as usual" ?
<Ariakenom>
what are nixos minimum ram requirements?
<exarkun>
Ariakenom: That depends on what you want to do with it.
<tilpner>
Ariakenom: Instantiating and building large systems can take multiple GB of memory, but it can run with less
<tilpner>
Ariakenom: So if you do those steps somewhere else, you can run with much less
<infinisil>
I'd get at least 8GB for a desktop system
<rnhmjoj>
dminuoso: well, you said the v2- commands don't work with nix but i didn't notice anything anusual when building inside a nix-shell created with cabal2nix. does that mean cabal was secretly using its own package set instead of nixpkgs?
<Ariakenom>
I was looking at the pinebook pro that has 4 GB and thinking ti sounded little
<dminuoso>
rnhmjoj: Well they work with nix, but not in an integration kind of way.
<dminuoso>
rnhmjoj: That is, you are not using any nix this way.
<dminuoso>
15:30:08 rnhmjoj | dminuoso: well, you said the v2- commands don't work with nix but i didn't notice anything anusual when building inside a nix-shell created with cabal2nix. does that mean cabal was secretly using its own package set instead of nixpkgs?
<dminuoso>
^- that last prat. Yes!
<dminuoso>
rnhmjoj: cabal v2-* maintains a global store in which packages are built into (using hashes to differentiate between different versions, flags... sound familiar?) - and then for each project, it reuses dependencies from that global store.
<dminuoso>
And equivalently, if new dependencies have to be installed, they are placed into the global store and referred to from the project.
<dminuoso>
Precisely what nix does with its nix store.
<qyliss>
Ariakenom: I think you could do fine with 4 GB
Chiliparrot has joined #nixos
<qyliss>
Shell: how much memory does your laptop have?
captn3m0 has quit [Read error: Connection reset by peer]
captn3m0 has joined #nixos
<infinisil>
Hm yeah 4GB should work, can still use swap if more is needed
<dminuoso>
rnhmjoj: So it really bypasses your entire nix stack, meaning you lose all the benefits of nix this way.
waleee-cl has joined #nixos
<dminuoso>
rnhmjoj: And that's why, if you want to use nix, you'd use `cabal2nix` - and then internally it ends up using cabal v1- behind the scenes.
<rnhmjoj>
dminuoso: ok, i understand now. i'm happy for other OS users they can share a bit of the nix joy but i'd like to keep using the real nix. do you know how to revert the previous behavior? i don't really like the v1- prefix
<dminuoso>
rnhmjoj: What do you mean?
<dminuoso>
rnhmjoj: You shouldn't be using v1- by hand. There's just so many things wrong with it.
<dminuoso>
In my opinion, use either nix with v1 or use v2.
jedai42 has quit [Ping timeout: 240 seconds]
<Shell>
qyliss: 2gb
<Shell>
qyliss: with 8gb swap
dreverri has joined #nixos
<rnhmjoj>
dminuoso: i meant using nix with v1-
<dminuoso>
rnhmjoj: Then you'd use cabal2nix - beyond that you dont invoke cabal anymore.
<dminuoso>
rnhmjoj: It's just raw nix.
<dminuoso>
rnhmjoj: cabal2nix will generate an appropriate haskell derivation that, underneath, does all the cabal magic. :)
<rnhmjoj>
dminuoso: ok, that i'd still use cabal to do stuff like generating the documentation, opening a repl..., no?
<qyliss>
Shell: do you often fill RAM?
<exarkun>
rnhmjoj, dminuoso: haskell.nix also supports cabal, though I've only used haskell.nix's stack support myself.
<Shell>
qyliss: not really tbh. right now I'm using about a gig. mind, I'm not using GNOME/etc.
<{^_^}>
[nixpkgs] @matthewbauer pushed 3 commits to master: https://git.io/Jea4t
<xantoz>
Ariakenom: [pinebook pro] should still be faster than the rpi3 I've experimented running nixos on (nixos-rebuild was pretty slow on that, though. had to disable documentation to make it run faster as it got stuck on a xsltproc invocation for a long time)
<LnL>
red[m]: we generally determine the target based on the number of packages that changed in the tree, your pr should have received a label with an estimation automatically
erasmas has joined #nixos
<red[m]>
10.rebuiold-linux: 501+
<red[m]>
?
<LnL>
yeah
domogled has joined #nixos
alexherbo2 has joined #nixos
alexherbo2 is now known as alex```
<kraem>
hey i'm curious as to how a package version is bumped. for example: linux-lts. there's a new version according to kernel.org (4.19.82) and linux-4.19.nix in the nixpkgs repo is stil at 4.19.81. i couldn't find any issues or pr for the bump to .82 - is it open to bump the version and create a pr?
<red[m]>
I have more work to do then to understand how all these branches hang together
kleschenko has quit [Ping timeout: 240 seconds]
<red[m]>
so, staging or staging-next?4
kleschenko has joined #nixos
<red[m]>
and then I guess I have to create a new branch off staging, make my changes and do another PR?
<jtojnar>
Changing the networkmanager modult to print `builtins.trace "Testing NM ${lib.boolToString cfg.enable}!!!!!!!!!!!!!!" cfg.enable`
<jtojnar>
I get both true and then false after each other when nixos-rebuild switch
<jtojnar>
why would that be?
drakonis has joined #nixos
zupo has joined #nixos
<dminuoso>
jtojnar: Perhaps that module is imported somewhere else? Im not quite sure how the evaluation model works with respect to several overlapping configs.
<red[m]>
#include <systemd.h8>
<thefloweringash>
Jan Tojnar: does passing `--no-build-nix` make the first one go away?
<red[m]>
At least on my servers, I really don't want another demon with the abilty to randomly change my network settings.
<red[m]>
I should probably go watch that systemd lightning talk that was referenced earlier on the migration to networkd to find out what's driving the want
drakonis_ has quit [Ping timeout: 240 seconds]
fendor has quit [Ping timeout: 250 seconds]
<jtojnar>
thefloweringash no, both are still there
bvdw has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @c0bw3b pushed commit from @r-ryantm to master « reaper: 5.983 -> 5.984 »: https://git.io/JeaR7
<worldofpeace>
manveru: the module is in https://github.com/NixOS/nixpkgs/pull/63813, but due to how fast upstream moves it'll be completely broken, so we need to upgrade the entire packageset first
<judson_>
I'm noticing (due to the glorious shellHook that lorri provides) that nix-shells will pass along their shellHooks (because they become part of the environment)
<judson_>
While I've found a workaround for using a nix-shell shebang within a nix-shell (`#! /usr/bin/env -S shellHook= nix-shell`) it seems like a bug.
<worldofpeace>
manveru: and it's really really difficult to figure out because they have so much software, I had a hard time parsing their changelogs so it may be a while before either romildo or I could do it.
<worldofpeace>
Jan Tojnar: what are the requested changes in https://github.com/NixOS/nixpkgs/pull/72860#issuecomment-550298269? I just wanted a good overview of the installed tests specifially, and having them under an attr `nixosTests.gnome-installed-tests.*` just made that easy. No relation to the gnome3 attr deprecation
<{^_^}>
[nixpkgs] @c0bw3b pushed commit from @r-ryantm to master « mcelog: 164 -> 165 (#72583) »: https://git.io/Jea08
<clever>
o1lo01ol1o: automatic gc is off by default, you need to manually run nix-collect-garbage as needed, or set min-free and max-free in nix.conf
<o1lo01ol1o>
clever: ah, ok thanks!
kleschenko has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dreverri has quit [Ping timeout: 268 seconds]
lordcirth_ has quit [Remote host closed the connection]
lordcirth_ has joined #nixos
<{^_^}>
[nixpkgs] @JohnAZoidberg opened pull request #72911 → vim_configurable: Add vi symlink to vim → https://git.io/Jea0a
<jtojnar>
worldofpeace what matrix client do you use? Fractal does not even allow copying multiple messages 😿️
<worldofpeace>
jtojnar: I jump from the electron app or fractal. I basically switch between either if something breaks. I actually copied them individually, does that not work?
selfsymmetric-mu has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #72899 → nixos/tests: Ignore shutdown/crash if not booted → https://git.io/Jea4J
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/Jea0D
<__monty__>
Anyone familiar with bonded interfaces and wpa_supplicant? http://ix.io/2100
<mightybyte>
infinisil: I'm getting errors when I try to do the nixos-rebuild switch.
<mightybyte>
And they're pretty inscrutable
kleschenko has joined #nixos
<mightybyte>
Active: failed (Result: exit-code)
<mightybyte>
This is using acme
<gchristensen>
__monty__: you probably can't use wpa_supplicant on a bond
<infinisil>
jtojnar: Ohh nice fix
<mightybyte>
acme-foo.mydomain.com.service: Failed with result 'exit-code'
<infinisil>
mightybyte: look at the full logs
<worldofpeace>
Jan Tojnar: lol, merging. With reviewers caught that before it also got backported to 19.09
<worldofpeace>
* wish
<jtojnar>
infinisil except for the infinite recursion
<infinisil>
journalctl -u domain.com -e
<infinisil>
mightybyte: ^^
noudle has joined #nixos
<infinisil>
jtojnar: Which inf rec?
<__monty__>
gchristensen: What makes you say this? I tested it when I set it up and it was switching between wired and wireless nicely then.
aveltras has joined #nixos
<jtojnar>
wireless.enable depends on delegatewireless which depends on wireless.enable
kleschenko has quit [Ping timeout: 245 seconds]
<jtojnar>
infinisil^
<gchristensen>
__monty__: my thinking is you probably need to use wpa supplicant on the bond slave instead of the bond itself
<gchristensen>
but I don't know
<__monty__>
gchristensen: Sure, but why would nix mess that up?
<gchristensen>
I'm not sure nix did
<gchristensen>
(did it?)
<infinisil>
jtojnar: Where do you get that? In the pr you just made?
shibboleth has joined #nixos
<mightybyte>
infinisil: Ahh, let me see...
<jtojnar>
infinisil yup
<infinisil>
Hm I see, maybe I'll take a closer look when I'm at home
<__monty__>
gchristensen: I *think* this is the relevant part of my configuration.nix: http://ix.io/2108
<infinisil>
,tell infinisil look at nm pr
<{^_^}>
infinisil: I'll pass that on to infinisil
mexisme has joined #nixos
<__monty__>
(Only switched to connman in the hopes of making the wireless more reliable. Switched to bonding because it still wasn't so I have to plug it in a lot.)
<detran>
what's the difference between vscode and vscode-with-extensions?
aveltras has quit [Quit: Connection closed for inactivity]
lsix_ has quit [Ping timeout: 245 seconds]
<hyperfekt>
detran: vscode-with-extensions means you manage the extensions in the nix expressions instead of them being taken from your user profile
<hyperfekt>
Can someone help me to compile qemu with debug symbols? I'm not being very successful. This is what I got so far, but gdb still complains about missing debug symbols: https://pastebin.com/09QuQ5gT
<detran>
thanks hyperfekt
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<tilpner>
hyperfekt: You're doing something weird to NIX_CFLAGS_COMPILE
<tilpner>
hyperfekt: Either it's a string before, and then you change it to a list
<tilpner>
hyperfekt: Or it's a list, then you have a nested list now
<hyperfekt>
tilpner: it's a string before and i make it a list. from my experiments either works?
<hyperfekt>
ooh or maybe it doesn't and it just doesn't give an error anywhere :/
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected. See: tofu-vim
<red[m]>
Question - I got asked to PR against staging which I have done. As I understand it changes here are batched before beinhg tested as they are typically for packages that have lots downstream
<red[m]>
Does my PR actually have to be merged-in to be tested?
drakonis has quit [Ping timeout: 268 seconds]
<red[m]>
(I'm not sure how the workflow is different relative to submitting PRs against master)
<red[m]>
okay- so kind of a pre-filter as it were?
<red[m]>
'sup otwieracz
<otwieracz>
I am working on derivation for BackupPC - and by some reason for every directory which installer is trying to create in installPhase I am getting errors like:
<otwieracz>
Failed to create or chown /nix/store/qn9k6x2bf4kk9v89j5kzjndvggbs9asq-backuppc-4.3.1/lib/Net
<hyperfekt>
something is stripping debug info from qemu and i can't figure out what. :< i got dontStrip = true and --disable-strip in the configureFlags :/
<karetsu>
I have however, run: `env XDG_CURRENT_DESKTOP=GNOME gnome-control-center` and when I try to click 'unlock' on the users after it shows no users I get an 'error acquiring permission' error - that be any use?
<clever>
no clue
<karetsu>
its some kind of dconf path I don't understand but if it thinks my user doesn't exist it makes sense that it wouldn't grant permission
<red[m]>
clever: any idea how I might go about debugging a nixos-rebuild build-vm that won't even boot?
<clever>
red[m]: is it giving an error?
<red[m]>
when I start it up it goes into a reboot-loop
<chessai>
clever: too many, but that's a different package. i'm just talking about the official arrayfire package, on the stable 3.6.4 release
<clever>
red[m]: if you know how to debug kernelsin gdb, you could probably do that, but it sounds like some pretty big breakage, can others reproduce it?
<chessai>
clever: to be clear, my PR is entirely independent of the haskell bindings
<clever>
chessai: ah
<red[m]>
oh, I may have a hypothesis
karetsu has quit [Quit: WeeChat 2.6]
<clever>
chessai: `strOnLinux = str: stdenv.lib.optionalString stdenv.isLinux str` could be reduced to `strOnLinux = stdenv.lib.optionalString stdenv.isLinux`
karetsu has joined #nixos
<clever>
chessai: all attributes of mkDerivation become env vars, so you could just directly `CUDA_PATH = cudatoolkit;`, and to make it optional, use `// lib.optionalAttrs ...`
<deni>
clever: This will disable a module altogether and I would have to copy paste the entire thing to change that one line. Right?
<clever>
deni: you would then copy the whole module file, edit the copy, and use imports to bring the copy in
emptyfl4sk has quit [Quit: WeeChat 2.6]
<red[m]>
I wonder if there is a qemu flag I can run to slwo the cpu down so I can see the errors before they reboot away...
<clever>
red[m]: you can disable rebooting entirely
<red[m]>
or mayeb I should jsyt mount the image and make certain that the kernel is actually there -P
<worldofpeace>
karetsu: can you do `sudo systemctl --runtime edit accounts-daemon.service` and insert `[Service] Environment=G_MESSAGES_DEBUG=all`?
<clever>
red[m]: build-vm uses -kernel on the cmdline, and no kernel actually exists in the image
<red[m]>
clever: I tried the no-reboot flag, and it just closes the windows or clears the screen on crash
<worldofpeace>
karetsu: (those should be on two lines btw)
<deni>
clever: ok that was my understanding as well. it's a shame that there isn't a better way if I just want to change a single line. the reason I want to change this is because the preStart script tries to chmod those directories and some of them (in my case) are a NFS mount so that fails and the service fails to start altogether
<worldofpeace>
karetsu: and then, log out log in, and hopefully you can paste entries from the journal from accounts-daemon with some info
<deni>
anyway thanks for the pointer
<karetsu>
worldofpeace: ok, brb
karetsu has quit [Client Quit]
arjen-jonathan has quit [Ping timeout: 250 seconds]
<clever>
deni: it would be best to modify the module to make that a proper option
<clever>
deni: you can use disabledModules to test the change, and then open a PR to nixpkgs to make it official
<chessai>
clever: not sure the optionalAttrs one is necessary, seems like style and less clear, at least to me
<worldofpeace>
hmm, it might actually have to be restarted. so if there's no info from systemctl on accounts-daemon.service karetsu you might have to manually restart it then log out
<chessai>
clever: i applied the other changes though
<clever>
chessai: yeah, its not really required
karetsu has joined #nixos
shibboleth has joined #nixos
<karetsu>
worldofpeace: I guess `systemctl status accounts-daemon` isn't enough? what's the journalctl arg I want?
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @JohnAZoidberg to master « apache-directory-studio: 2.0.0.v20170904-M13 -> 2.0.0-M14 »: https://git.io/Jeawm
<jd823592>
I have got services.emacs.enabled = true and I don't remember how my .config/systemd/user/emacs.service got to link to my emacs-26.1 installation... now i upgraded emacs, ran nixos-rebuild switch and got 26.3 but the systemd unit keeps running 26.1, any ideas
<chessai>
jd823592: how did you upgrade emacs?
<karetsu>
worldofpeace: it just says that its starting and stopping the accounts service, I'll reboot
karetsu has quit [Quit: WeeChat 2.6]
<worldofpeace>
karetsu: those edits you made to the service don't persist
<jd823592>
chessai: I tried bunch of things, nixos-rebuild switch --upgrade, nix-channel --update, nix-env -u emacs, ...
<chessai>
`nix-env` wouldn't affect a service
karetsu has joined #nixos
<jd823592>
i figured that
lovesegfault has quit [Ping timeout: 264 seconds]
<worldofpeace>
karetsu: hmm, maybe we have to go the route of dbus-monitor on a tty
<jd823592>
i see system-path among referrers to emacs-26.1 :/
<karetsu>
yeah, it showed nothing beyond starting and stopping
lovesegfault has joined #nixos
<chessai>
jd823592: so you didnt actually upgrade emacs. and, did you run `sudo nix-channel --update`?
<karetsu>
worldofpeace: I have no idea what the type, sender or interface need to be for that in dbus-monitor though - just monitor it all and only do a log in?
<jd823592>
chessai: i did, and I most certainly have 26.3 in my user environment, why didnt nixos-rebuild pickup that update too?
<chessai>
jd823592: define 'user environment'
karetsu has quit [Client Quit]
<jd823592>
my regular user (uid=1000): `nix-env -q` contains emacs-26.3
<clever>
jd823592: what was the path to the emacs .service file?
<clever>
jd823592: the service runs whatever you put into services.emacs.package
<clever>
jd823592: so you can just set services.emacs.package = unstable.emacs;
<karetsu>
worldofpeace: I should also note that nuking /v/l/A/users never ever regenerated them, I'll try to add them in manually
kleschenko has quit [Ping timeout: 264 seconds]
kleschenko has joined #nixos
<worldofpeace>
karetsu: ?
<worldofpeace>
karetsu: deleting the individule files in that dir never created new one's for each user?
<karetsu>
nope
lsix_ has quit [Ping timeout: 252 seconds]
soju has joined #nixos
<karetsu>
and there's still none there after a clean install, if I can get what they should look like and me just writing them patches this annoyance I'll take it
<worldofpeace>
karetsu: and your logs log completely uneventful, locally I tried doing the same steps using gdm and I can see events for the gdm user and my user etc.
ng0 has joined #nixos
<worldofpeace>
karetsu: what's a clean install? Does this mean reformat?
<karetsu>
yes
<karetsu>
I am currently on a completely reformatted machine
<karetsu>
and still no dice
<karetsu>
it never generated the /var/lib stuff to begin with
<karetsu>
wait, restarted and it sees /etc/shadow again
<karetsu>
I am so confused rightnow
<worldofpeace>
karetsu: I see nice g_debug in that source, did editing the service at runtime to include G_MESSAGES_DEBUG=all really include no logging?
<karetsu>
it added none, I'll do it again to prove I'm insane :)
<karetsu>
it was under the [Service] header right?
<karetsu>
and is it ok that its giving it some heinous file name?
<worldofpeace>
yep, [Service] and Environment=
<worldofpeace>
if for whatever reason it doesn't work, maybe you can run the accounts-daemon on a tty with --debug
jd823592 has quit [Remote host closed the connection]
<worldofpeace>
karetsu: that's just systemd being systemd
<karetsu>
ok then , time to log out and in
karetsu has quit [Quit: WeeChat 2.6]
<worldofpeace>
you need to restart it before that I think
<karetsu>
worldofpeace: should I restart it as well or just log out and in? because its added 0 entries to journalctl -u accounts-service
<worldofpeace>
karetsu: I believe sudo systemctl restart accounts-daemon.service. and then logout
<worldofpeace>
else fails you can stop the service and switch to a tty, doing `nix build nixpkgs.accountsservice; .sudo ./result/libexec/accounts-daemon --debug`
karetsu has quit [Client Quit]
karetsu has joined #nixos
karetsu has quit [Client Quit]
o1lo01ol1o has quit [Remote host closed the connection]
karetsu has joined #nixos
<{^_^}>
[nixpkgs] @tohl opened pull request #72953 → xfburn: 0.5.5 -> 0.6.1, tested on nixos x86_64 → https://git.io/JearT
<worldofpeace>
ooh, here's the dbus-send command you can provide at a later step to see all the properties accountsservice knows https://gist.github.com/worldofpeace/152dab9d16c7624bd00edcf7af5184fa. but we're not at that step yet, I'd just want to see accounts-daemon running with some sort of debug output. Note that User1000 may be User$uid as your current user id may be different.
<worldofpeace>
are you supplying the method input karetsu ? I'd expect it to return `no such interface properties` if you're missing string:"org.freedesktop.Accounts.User"