eyenx has quit [Remote host closed the connection]
vs^ has joined #nixos
<infinisil>
Opinion: Every somewhat complicated Nix code should have an option to explicitly log what it does
<bqv>
infinisil: there should definitely be multiple, as a rule of habit I make at least both ed25519 keys and rsa keys on every server/user
<V>
Opinion: Nix should have an interactive debugger :p
<bqv>
infinisil: V: both solved by people using options more instead of let!
<infinisil>
Yeah, interactive debugging would be awesome, but in absense of one, logging is probably the next best thing
<infinisil>
bqv: *confused*
ardumont has quit [Ping timeout: 264 seconds]
eyenx has joined #nixos
<energizer>
a --trace option that logs everything
<infinisil>
Okay so, we should also have different logging levels
jeaye has joined #nixos
hoijui has quit [Ping timeout: 244 seconds]
<bqv>
No, I mean, complicated nix code is less complicated if you can introspect it at points (options), and your repl becomes an interactive debugger if you can introspect eval at any point (options!)
<infinisil>
Maybe a `builtins.log { verbosity = 10; component = [ "pkgs" "hello" ]; message = "Evaluating hello with ..."; }` would be nice
fuiltilt has joined #nixos
<bqv>
I use trace to almost that effect, for key stuff, yeah
ardumont has joined #nixos
<infinisil>
bqv: Ah i see yeah
<infinisil>
Though there's places where the module system can't be used
<infinisil>
Like, for all package definitions (because it would be too slow)
<infinisil>
Or for the module system implementation itself :P
<bqv>
infinisil: ha, I was thinking of trying that this morning (re pkgs), good to know.
<bqv>
infinisil: ah, cool
<tpw_rules>
how can i "package" a nix-shell? i want to be able to put some derivation in my environment.systemPackages or whatever, type a command, and get put into that shell environment
<bqv>
Well now that nixus is irreversibly my de-facto deploy system maybe I'll do the ssh thing...
<infinisil>
bqv: Ayyy, let me know if you have questions :D
tad-lispy[m] has joined #nixos
<infinisil>
tpw_rules: Hmm interesting..
<tpw_rules>
my use case is i have an internal software system that's a number of different programs, along with a python interpreter with a bunch of packages. so i want to be able to install the "system" as its own thing, then run a command to get into a shell and run tool_a, tool_b, python -c "import tool_c" etc. plus this would also keep it from getting garbage collected
<remexre>
is there any version of nix for i686 darwin?
<remexre>
slash any hope of easily getting it set up
<infinisil>
It's not perfect, there's some weird escape issue, and if you check the file it sources, there's quite a bunch of environment variables you probably don't want to set. But it kind of totally works
<KarlJoad>
abathur: I mean that the one writing a new package (for another package for Octave) can just say `patchPhase = ''some commands'';`. But when they don't use `patchPhase` at all, `stdenv.patchPhase` is used instead.
ransom has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<abathur>
I can't really recall this well enough to explain it off the top of my head, but I managed to distill it out of the code for the python builder
fryguy has joined #nixos
<KarlJoad>
So, I essentially _have_ to perform a set-union to get the behavior I want?
<abathur>
IIRC you can ignore most of the middle; you need to look at the function args up top, with the @attrs:, and then start paying attention again at "self = (stdenv.mkDerivation"...
tgamblin-llnl has joined #nixos
<abathur>
I think so; I don't really know so
<abathur>
I found that example, and did it that way, and no one barfed on me during PR review :)
<KarlJoad>
Fridh mentioned something about that in my PR, so I guess that's the way to do it.
<abathur>
I think extend and @attrs are both also playing a role here, but it's been most of a pandemic since I did the initial draft of that bit
<abathur>
*lib.extendDerivation
Ashy has joined #nixos
<KarlJoad>
Gotcha. I mean, in theory, a // is what I want. I just don't know how to phrase the rest of what I want yet. I have to type it out and test it.
rajivr has joined #nixos
<Ashy>
i've mounted a samba share like so: https://nixos.wiki/wiki/Samba#cifs_mount - how do i mount it so my normal user account has read/write access to the mount point?
bridge[evilred] has quit [Ping timeout: 260 seconds]
sarcasticadmin has quit [Read error: Connection reset by peer]
<dmj`>
I've written a derivation for a vpn client (cisco anyconnect vpn). It has a lot of hard-coded calls to open files in /opt/cisco/anyconnect, I use makeWrapper and libredirect to try and get around this. The sad part is that this doesn't seem to be working. When I call strace the calls to open files in /opt/cisco/anyconnect are still failing. Anyone know why this would be the case?
sarcasticadmin has joined #nixos
tgamblin-llnl has quit [Ping timeout: 265 seconds]
<bqv>
infinisil: your dns multimod... is that a dns server?
zebrag has quit [Quit: Konversation terminated!]
<bqv>
dmj`: libredirect isn't perfect
<bqv>
your calls are probably leaked
<bqv>
you could test that by creating the missing files and see if it works then :p
<dmj`>
bqv: what do you mean by leaked ?
<bqv>
libredirect just overrides a specific syscall
<bqv>
that is one out of hundreds of getting files
<dmj`>
bqv: I have copied the missing files into $out, and used NIX_REDIRECT to point /opt/cisco/anyconnect=$out
waleee-cl has quit [Quit: Connection closed for inactivity]
<DigitalKiwi>
how do i keep kwin from starting when i start a plasma5+xmonad session
<bqv>
friends
<bqv>
i'm thinking of making my config depend on builtins.exec be available
<bqv>
convince me otherwise?
<DigitalKiwi>
do it
<bqv>
dammit guys
m4ts has joined #nixos
<bqv>
see, my thinking is, guix doesn't have this distinction between eval/build/execute, and i don't fully understand why nix makes it an absolute requirement
<bqv>
anonymous metrics would be an idea, but stuff in the nix world moves at a glacial pace, i'd prefer to not waste time on things that aren't key
nolash has quit [Ping timeout: 256 seconds]
respawn_ has quit [Quit: Leaving]
ml| has joined #nixos
<DigitalKiwi>
but i want to know how many people use the programs i spend countless hours making work :(
nolash has joined #nixos
<DigitalKiwi>
it'd be nice if plasma5 worked too :(
<DigitalKiwi>
still haven't fixed...
hnOsmium0001 has quit [Quit: Connection closed for inactivity]
<DigitalKiwi>
why does it even install wayland
<DigitalKiwi>
i can't use wayland
<DigitalKiwi>
it's called Xmonad not WaylandMonad
mallox has joined #nixos
<bqv>
DigitalKiwi: yeah that'd be possible, but again, would you rather have random metrics, or less nix bugs and more nix features
<DigitalKiwi>
why not both
<bqv>
cause dude look at the PR list, eelco just ignores most of them so there's more than i can count rn that are super cool but aren't going anywhere, and he can only dev so fast on his own
<bqv>
and let's not even talk about nixpkgs
<bqv>
so yeah, both's not really happening
<DigitalKiwi>
who said anything about eelco
<bqv>
well how else you gonna get it merged in
<DigitalKiwi>
??
chour has joined #nixos
<bqv>
you want it in nix, presumably, cause otherwise it's nixOS metrics, not nix metrics, and more people use nix and not nixos, than nixos
<DigitalKiwi>
same way i get any program put in nixpkgs
zakame has quit [Ping timeout: 272 seconds]
Mateon1 has quit [Ping timeout: 240 seconds]
<DigitalKiwi>
why's it have to be part of nix?
<bqv>
so i download nix on my ubuntu machine, install a load of stuff and home-manager, use nix extensively, but cause i'm not nixos i'm not in your metrics
<kini>
In a nixpkgs checkout, `nix-build . -A freetype.dev` builds `/nix/store/...-freetype-2.10.4-dev`, but `nix build -f. freetype.dev` builds `/nix/store/...-freetype-2.10.4`. Is this expected?
<cole-h>
kini: result-dev will have the dev output
<kini>
Right, but how can I get it using `nix` (rather than `nix-build`)?
<kini>
oh, never mind, I see. derp
<cole-h>
Yep :)
<kini>
you mean the symlink literally named `result-dev` will point to the right output :)
<cole-h>
I don't know why, but `nix build` builds all outputs, even when one is specified
<cole-h>
¯\_(ツ)_/¯
<kini>
Oh well, thanks for mentally doing `ls -l` for me :P
<cole-h>
:D
<kini>
Actually, it looks like `nix build -f. freetype.dev` only produces a `result-dev` symlink, at least on my version of `nix` (2.3.10). The `result` symlink must have been left over from when I had run `nix build -f. freetype` previously...
sangoma has joined #nixos
<cole-h>
Not so, at least on my machine. I get both even after unlinking `result` and `result-dev`
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #109060 → nixos/xdg/portal: fix loading portals from systemd service → https://git.io/Jtvfr
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JtqSl
asheshambasta has left #nixos ["ERC (IRC client for Emacs 28.0.50)"]
simba1 has quit [Ping timeout: 244 seconds]
ahmedelgabri has joined #nixos
lsix1 has joined #nixos
ahmedelgabri has quit [Client Quit]
<{^_^}>
[nixpkgs] @spease opened pull request #110106 → Enable halide on x86_64-darwin → https://git.io/JtqSw
m0rphism has joined #nixos
cole-h_ has joined #nixos
puffnfre- has joined #nixos
adisbladis has quit [Ping timeout: 256 seconds]
puffnfresh_ has quit [Ping timeout: 256 seconds]
m4ts has quit [Remote host closed the connection]
sarcasticadmin has quit [Read error: Connection reset by peer]
sangoma has quit [Ping timeout: 256 seconds]
ml| has quit [Ping timeout: 256 seconds]
cole-h has quit [Ping timeout: 256 seconds]
cransom_ has quit [Ping timeout: 256 seconds]
cransom_ has joined #nixos
sangoma has joined #nixos
zakame has quit [Ping timeout: 256 seconds]
cr4y1 has quit [Ping timeout: 256 seconds]
otulp has quit [Ping timeout: 256 seconds]
statusfailed has quit [Ping timeout: 256 seconds]
m4ts has joined #nixos
otulp has joined #nixos
statusfailed has joined #nixos
zakame has joined #nixos
cr4y1 has joined #nixos
ml| has joined #nixos
adisbladis has joined #nixos
<{^_^}>
[nixpkgs] @LeSuisse opened pull request #110107 → cifs-utils: missing python3 dependency for smbinfo → https://git.io/JtqSb
<pinpox>
Hi, nixos-rebuild stated throwing an error: "daemon at '/nix/var/nix/daemon-socket/socket': Connection refused" what does that mean exactly and how can I fix it?
<illustris>
When running make, it fetches dependencies using curl. As far as I understand, nix builds are not allowed to make network calls in the build stage, causing this to fail. Is there anything I can do to get around this?
<esotericn>
am I able to use callPackage inside pythonPackages.override? i basically want to add my own packages into the pythonPackages set as an overlay in a sane way, similar to how nixpkgs does it
<esotericn>
passing self/super through as arguments seems janky
<infinisil>
bqv: The dns multimod starts bind yeah, though I want to add support for other dns servers too (maybe also with a backend option like with vpn)
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtmLw
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 0 commits to emplace-0.4.2: https://git.io/JtmLo
johnnyfive has quit [Quit: I quit now.]
johnnyfive has joined #nixos
ahmedelgabri has quit [Ping timeout: 260 seconds]
<itai33[m]>
i have a local repo of mpd with some modifications, and i want to use nix to build it. is there a way i can just copy paste the default.nix from nixpkgs and do it?
<djanatyn>
itai33[m]: you could copy your changes to the repository as a patch (git diff -p), or you could push your changes to a fork and override the `src` attribute of the derivation: https://nixos.org/manual/nixpkgs/stable/#sec-pkg-overrideAttrs
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 57 commits to staging-next: https://git.io/Jtmta
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 58 commits to staging: https://git.io/JtmtV
<djanatyn>
if you're in your git checkout, you can do something like this: nix repl /var/lib/nixpkgs <<< ':b mpd.overrideAttrs (oldAttrs: rec { src = ./.; })'
<itai33[m]>
djanatyn: but like isn't there a way to have a `nix.shell` that you can use to build the project locally or something?
<{^_^}>
[nixos-homepage] @garbas pushed 73 commits to experiment-with-different-fonts: https://git.io/JtmtH
<itai33[m]>
let me try that
<djanatyn>
itai33[m]: sure
<djanatyn>
replace /var/lib/nixpkgs with your nixpkgs checkout, or a channel
<itai33[m]>
djanatyn: i can't seem to find the path where the channel is saved
<itai33[m]>
or should i like put in a url
<djanatyn>
i think you can do this: nix repl '<nixpkgs>'
<djanatyn>
assuming you have nixpkgs on $NIX_PATH
<itai33[m]>
yep that works
<itai33[m]>
still though, i would rather just have a `nix.shell` so i can just do `nix-shell` and have it just work
<itai33[m]>
cause i feel like i should be able to do that i just don't know enough nix yet
<djanatyn>
you can :)
<djanatyn>
once you have a working nix expression, you can drop that in a file shell.nix and then `nix-shell` can load it
<djanatyn>
i usually experiment to get the right expression using `nix repl`, and then move that expression into a file (default.nix or shell.nix) once it's working
<djanatyn>
nix repl has some convenience functions too. it can install to local profiles, instantiate nix-shell, build dependencies, etc: nix repl <<< ':?'
<itai33[m]>
I see
<itai33[m]>
what does :b do?
<djanatyn>
you can check with `:?`, it just builds a derivation
<djanatyn>
"You can also override[1] your package in a shell.nix file to add test and coverage dependencies, that are not necessary for the actual build of the package, but that you want for your development environment." https://nixos.wiki/wiki/Development_environment_with_nix-shell
lsix1 has joined #nixos
<itai33[m]>
djanatyn: wait it just built the program and then dropped me back into the shell
<itai33[m]>
like zsh not the nix shell
<djanatyn>
yeah
<itai33[m]>
thanks for the link i'll read it
<itai33[m]>
oh right cause `:b` stands for build
<djanatyn>
you could replace `:b` with `:s` to drop to nix-shell, or you could do it more directly: nix-shell -p 'mpd.override (oldAttrs: rec { src = ... })'
<djanatyn>
lots of approaches will work
<itai33[m]>
when i try to add the `mpd.override..` into the buildInputs of shell.nix it gives me a "cannot coerce a function into a string" error
<itai33[m]>
maybe i should surround it with curly braces or somehting?
<djanatyn>
`--show-trace` usually gives more context for errors, you can also pastebin your derivation
<bqv>
Why does nixpkgs have services.wg-quick and services.wireguard
konobi has quit [Remote host closed the connection]
<stefandeml>
Hello everyone, quick question: I'm trying to automate nix installation via an installer script / ansible. However the provided installer uses so UI confirmation and afaik there is no way to run it headless? - headless check seems to be only based on asserting fd 0 with a terminal? - any ideas on how to solve this? - Thanks!
konobi has joined #nixos
<stefandeml>
I also think in the past there have been some deb packages for nix? - are they still provided?
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtmsL
mmmattyx has joined #nixos
<bqv>
infinisil: wow, I think you can legit just replace wg-quick with wireguard and it should still work. I'm very confused why there are to structurally isomorphic ways to do the same thing, at least in nixpkgs
<supersandro2000>
stefandeml: debian has packages which are good enough for boot straping
<supersandro2000>
make sure to turn on sandbox though
zebrag has joined #nixos
xeronull has joined #nixos
zebrag has quit [Client Quit]
<infinisil>
bqv: But like for nixus, I don't think it makes sense to support both wg-quick and wireguard, when the former seems to be just a strict superset in functionality
<bqv>
infinisil: I just struggle to justify using a wrapper when there's actually no need, I'd rather not, but understandable if you want to keep it, since I'm subtree forked I can modify it locally so long as you don't change your mind later :p
poscat has quit [Client Quit]
<infinisil>
bqv: For me, wireguard on its own doesn't work, because I want to route all my traffic through the server
<infinisil>
I'm not sure how I could make it work with wireguard on its own, but wg-quick takes care of it automatically
magnetophon has joined #nixos
poscat has joined #nixos
NieDzejkob has quit [Read error: Connection reset by peer]
<sss2>
hi all, what difference between nixos and nixpkgs channels, can i use nixos channel on foreign host ?
<dutchie>
nixos channels have some extra/different tests run on them before the CI lets them advance
<sss2>
so i can use it on foreign host ?
<dutchie>
no reason you couldn't use it on a foreign host, but sometimes the nixos channels lag behind a bit because something broke one of the tests that you don't care about
<sss2>
hmmm
xeronull has quit [Quit: Connection closed]
<sss2>
i have read what -small have latest packages, but here is not nixpkgs-small
<ixxie>
possibly related issue - nix-prefetch-github google fonts seems to stall
lordcirth_ has joined #nixos
<{^_^}>
[nixpkgs] @charvp opened pull request #110181 → aerc: replace an extra hardcoded path → https://git.io/Jtm8I
orivej has quit [Ping timeout: 256 seconds]
sangoma has joined #nixos
KarlJoad has joined #nixos
<lukegb>
if the sha256 is wrong but matches e.g. a different revision, then it'll probably still work
<lukegb>
iirc the rev isn't an input to the hash function, so if you change the rev without changing the hash it'll just use the old content
stefandeml has quit [Quit: Ping timeout (120 seconds)]
<KarlJoad>
I'm getting an interesting issue, where the patchPhase I wrote isn't being run when `dontPatch = false`. I'm not sure why it's happening at all.
nixuser_ has quit [Remote host closed the connection]
<kunrooted[m]>
okie, it's been a while since I've lastly asked for help in here
<kunrooted[m]>
but I wanted to ask about something, my other Buddy mentioned flakes, do they change the config very much? or not? just asking, may want to start using them in the future too
<kunrooted[m]>
I've started using that home-manager, and I have pretty basic config and it works so I'm pretty happy about it, I even told one of my Buddies about nixOS and the greatness of it and he started using it too so now we're sharing a config haha
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #110184 → ocamlPackages: irmin 2.2.0 → 2.3.0; git: 2.13 → 3.1.0; and related changes → https://git.io/JtmRu
mmmattyx has quit [Quit: Connection closed for inactivity]
erasmas has joined #nixos
<kunrooted[m]>
not yet, started reading wiki tho
<kunrooted[m]>
it doesn't change the config a lot, or does it?
orivej has quit [Ping timeout: 256 seconds]
simonpe^^ has joined #nixos
<simonpe^^>
I'm trying to do some querying on my CI expressions with `nix-instantiate --json --strict --eval -E 'with import ./. {}; (sources.hello-world2.devices or [])'` but I'm getting `error: cannot import /nix/store/xxxxxxxx/sources.nix', since path '/nix/store/yyyyyy.drv' is not valid`, I'm assuming nix-instantiate isn't instantiating whatever is required?
<azazel>
I'm not a home-manager user, but flakes are a new way of packaging nix stuff, so maybe how you reference home-manager sources and you own configuration changes, but the configuration per se shouldn't change
<simonpe^^>
and the drv doesn't exist
<kunrooted[m]>
hmm, got it
<simonpe^^>
And I realize that if I remove the --eval flag it will instantiate the correct stuff but then it doesn't print the output of my query
<{^_^}>
[nixpkgs] @fgaz opened pull request #110186 → Twemoji color font/dont build from source → https://git.io/Jtm0e
<azazel>
simonpe^^: that behaviour is as expected, as --eval flags to not create the derivation
<simonpe^^>
can I get the behavior I want without running the command twice, once with --strict and once with --eval?
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #110187 → ocamlPackages.curly: add missing meta set → https://git.io/Jtm0s
euandreh_ has joined #nixos
<kunrooted[m]>
got an answer from someone, seems like it can change a lot
<bqv>
Neat, they fixed the xonsh bug
<kunrooted[m]>
but it seems like a really cool idea to try it one day
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @mupdt to master « terraform-providers.shell: provider source address (#110160) »: https://git.io/Jtm09
aveltras has quit [Quit: Connection closed for inactivity]
pitch has joined #nixos
malook has joined #nixos
Lord_of_Life_ has joined #nixos
Lord_of_Life has quit [Ping timeout: 246 seconds]
cybersyn has joined #nixos
Lord_of_Life_ is now known as Lord_of_Life
<Cadey>
i have a list of attrsets, how do i make it into a larger composite attrset?
ahmedelgabri has quit [Ping timeout: 260 seconds]
ahmedelgabri has joined #nixos
<simonpe^^>
Cadey: stdenv.lib.recursiveUpdate and a fold
ManiacOfMadness has joined #nixos
ahmedelgabri has quit [Ping timeout: 240 seconds]
rb2k has quit [Read error: Connection reset by peer]
rb2k has joined #nixos
meh` has quit [Ping timeout: 264 seconds]
<slby[m]>
<kunrooted[m] "https://github.com/hlissner/dotf"> Would it be the NixOS experience if I gave you all the answers in one, convenient place? :D
<kunrooted[m]>
this haha
<kunrooted[m]>
tho it's possible, I'd say
<kunrooted[m]>
I have all answers here and on #nix:matrix.org
hiro98 has quit [Quit: hiro98]
hiro98 has joined #nixos
<KarlJoad>
I'm getting an interesting issue, where the patchPhase I wrote isn't being run when `dontPatch = false`. I'm not sure why it's happening at all.
<anderscs[m]>
Does anyone have good tips on building go/nodejs/electron apps? I'm trying to compile the itch.o desktop app via the itch-setup installer https://github.com/NixOS/nixpkgs/pull/110012
<{^_^}>
#110012 (by Anderssorby, 21 hours ago, open): itch: init at version v1.24.0
<fast_invsqrt[m]>
I'm trying to add a Romanian dictionary to NixOS. I am doing this by editing the dictionaries.nix in `pkgs/development/libraries/hunspell/`. Is this how I'm supposed to be doing it?
<fast_invsqrt[m]>
* I'm trying to add a Romanian dictionary to NixOS. I am doing this by editing the `dictionaries.nix` in `pkgs/development/libraries/hunspell/`. Is this how I'm supposed to be doing it?
evanjs has quit [Ping timeout: 240 seconds]
hoijui_ has quit [Ping timeout: 272 seconds]
hoijui has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @milahu opened pull request #110197 → firewall: move rpfilter to mangle.PREROUTING to fix conntrack → https://git.io/JtmaA
<codygman__>
All of a sudden `stack --nix` errors with 'attribute ghc8103 missing' even though I demonstrate in the following paste ghc8103 exists, any ideas? https://pastebin.com/Z1Qdw0UV
<energizer>
does `unpacking 'https://github.com/...` mean "downloading and unpacking" or just unpacking?
<avn>
Folks, anyone use npmlock2nix by tweag? It works well for me on linux, but create problems on darwin
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 85 commits to staging-next: https://git.io/Jtmr2
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 86 commits to staging: https://git.io/Jtmra
<lordcirth>
avn, you should say what the problem is up front
<avn>
probably ;)
<avn>
node-gyp raising `libtool: error: unrecognised option: '-static'` error. With node2nix it was handled with adding pkgs.xcodebuild, but here it not helps
<lordcirth>
avn, are you sure you are using the same versions of everything on darwin? What channel are you using?
<{^_^}>
[cabal2nix] @peti pushed to ci « Run the test suite in C.UTF-8 locale rather than en_US.UTF-8. »: https://git.io/Jtmrp
<{^_^}>
[nixpkgs] @primeos merged pull request #110198 → nixos/tests/chromium: Add ungoogled-chromium and Google Chrome → https://git.io/JtmVQ
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<lordcirth>
avn, same version of Nix itself?
CodeWarrior has joined #nixos
<KarlJoad>
I'm getting an interesting issue, where the patchPhase I wrote isn't being run when `dontPatch = false`. I'm not sure why it's happening at all.
<Nevoic>
Hey! I was wondering if I could have some help setting up an android project using a shell.nix file. I'm unable to get licenses to accept.
<Nevoic>
I have `config.android_sdk.accept_license = true;` set, and I've tried a couple different variations of it and nothing seems to work.
ahmedelgabri has joined #nixos
<Nevoic>
If I could just run some kind of post install script to modify the nix store that would work too, since all I really need to do is run `yes | sdkmanager --licenses`, but since the nix store isn't modifiable, this ends up doing nothing.
<{^_^}>
[cabal2nix] @peti pushed to ci « Try to guess proper doctest cmdline flags to make the tool succeed. »: https://git.io/Jtm6c
<Nevoic>
I have also tried `{ pkgs ? import <nixpkgs> { config.android_sdk.accept_license = true; } }:` on line one, and removing the `pkgs =` within the let.
<Nevoic>
but it behaves exactly the same way.
<Nevoic>
Can anyone see my messages? I'm not exactly sure if this IRC client is working.
<Nevoic>
I also tried putting it at the top of the file like `{ pkgs ? import <nixpkgs> { config.android_sdk.accept_license = true; } }:` and removing the `pkgs = ` line
<Nevoic>
but they both behaved in the same way.
<clever>
those would behave identically
cybersyn has quit [Ping timeout: 256 seconds]
<clever>
development/mobile/androidenv/compose-android-packages.nix: by setting nixpkgs config option 'android_sdk.accept_license = true;'
<clever>
development/mobile/androidenv/default.nix:, licenseAccepted ? config.android_sdk.accept_license or false
<clever>
i would expect that to just work
<energizer>
sorry this is silly, i can't find the function for this. i have [{name = "a"; foo = 1; } {name = "b"; foo = 3; }] and I want {a = {name = "a"; foo = 1; b = {name = "b"; foo = 2;}}
nickywhite[m] has joined #nixos
<energizer>
s/3/2/
<Nevoic>
clever so you're saying what I did should work?
<clever>
Nevoic: i would expect it to just work
<Nevoic>
the error I get is "License for package Android SDK Platform 29 not accepted".
<clever>
Nevoic: i think that error is a different one, not the android_sdk.accept_license one
<Nevoic>
I saw something about `extraLicenses`, but I couldn't set it.
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JtmPH
<Nevoic>
Yeah, I have a line like `unstable = import <nixos-unstable> { config = { allowUnfree = true; }; };` in my configuration.nix, I assume the same kind of thing would work in a local shell.nix?
<Nevoic>
because I've had to install unstable packages before.
<clever>
Nevoic: yep
<clever>
Nevoic: you could also use something like niv, to pin your shell.nix to a specific nixpkgs rev, and it wont update on its own
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtmXT
<energizer>
pistache: yes
<{^_^}>
[nixpkgs] @peti pushed 227 commits to haskell-updates: https://git.io/JtmXI
<pistache>
energizer: I want to build a Python package, while adding a file (it's a plugin) to the package itself
<Nevoic>
clever: so how would I figure out what license is specific to Android SDK platform 29?
<{^_^}>
[nixpkgs] @SuperSandro2000 opened pull request #110210 → Fix collection 4: substituteInPlace cleanups, pythonPackages.azure-synapse-artifacts: Fix meta → https://git.io/JtmXL
<asbachb>
Can I regenerate hardware.nix somehow?
<Nevoic>
the unstable stuff worked! but the config option still doesn't work, which I think is to be expected.
<pistache>
the package in question has a "plugins/" folder, I wrote a custom plugin, and I'd like to make a derivation "pkg-with-my-plugin" that includes that file in the "plugins/" folder
<clever>
Nevoic: not sure, ive not done android in a while
<pistache>
the unit of code reuse is indeed a package, but some packages abuse this with the implementation of their plugin systems
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pistache>
the package in question is "folium"
evanjs has quit [Read error: Connection reset by peer]
<KarlJoad>
asbachb: `nixos-generate-config`, like you did when you did your first install, should generate that again, I think.
davidv7 has joined #nixos
<energizer>
pistache: how does the folium plugin system work?
<pistache>
energizer: it looks for Python modules in its "plugins/" subdirectory
<energizer>
pistache: like, literally?
<energizer>
pistache: or it imports from folium.plugins
<pistache>
sorry, it imports from folium.plugins
<{^_^}>
[nixpkgs] @symphorien merged pull request #109894 → sysklogd: fix building against musl → https://git.io/JtLbk
<energizer>
pistache: are there docs on how to write your own folium plugin?
kalbasit has quit [Ping timeout: 265 seconds]
lsix1 has joined #nixos
<pistache>
energizer: not that I know of
<KarlJoad>
I'm getting an interesting issue, where the patchPhase I wrote isn't being run even when `dontPatch = false`. I see "patching sources", but the `sed`s in patchPhase don't have any effect.
<Nevoic>
clever: hmm, extraLicenses still isn't working. I decided to just entirely replace `pkgs` with the unstable version in my shell.nix to make sure I wasn't missing changing something from 20.09 to unstable, and it's still not an available configuration option.
<Nevoic>
is it possible this commit is just on master and not unstable yet?
<energizer>
pistache: are you sure they *have* a plugin system?
<clever>
Nevoic: pretty sure unstable has updated in the last 2 months
<energizer>
pistache: you can just put it in postPatch or whatever but i think that's unnecessary magic. you can just import your code from wherever it actually lives instead of monkeypatching folium
<pinpox>
I'm trying to reinstall the bootloader because my BIOS can't find it. (Dualboot with windows, only windows shown). I ran "nixos-rebuild --install-bootloader switch" and it didn't throw any errors, but I still can't boot into my system without the help of a installation medium's bootloader
<colemickens>
I know there's a cli branch for nix cli but this is nice in the meantime
<V>
doesn't combine nicely with nixos-rebuild
<V>
and also is wonky if you're building on remote systems
<quinn>
abathur: you're right
meh` has joined #nixos
<pinpox>
I switched from an older cpu+mobo to a b550+ryzen, anything special I have to do to get nixos to boot? The installer boots fine
<abathur>
quinn it doesn't look like ofborg ran the equivalent darwin check on the init PR a few weeks ago, so I guess it's possible there's an actual hang in there and this is just the first chance to find it
<maralorn>
colemickens: I never dared to announce it because it didn’t feel ready enough.
<colemickens>
pinpox: did you mount /boot before running the command you mentioned?
<pinpox>
oh maybe I forgot that. Let me run that agin
<pinpox>
it's lvm+luks and flakes, in case any of those make any difference
<maralorn>
V: What is wonky about using remote systems?
<quinn>
abathur: on the main commit page, it also just says "waiting to run this check"... very mysterious. i also would have expected a timeout by now
<colemickens>
pinpox: fwiw you'll probably want to make sure you're booting the install media via UEFI so that efibootmgr can do its job.
<V>
maralorn: it wants to poke around in the nix store, and if it's looking in your local nix store while things are building on another machine, you won't get very useful output
<abathur>
I think 12h may be the timeout, but I don't remember why I think that
<madmaurice>
KarlJoad, it builds for me. I also see "patching sources". Do you by any chance have the same package within an overlay?
<V>
maralorn: I think you can repro trivially with nixos-rebuild --target-host <...> switch
<quinn>
abathur: alright, i'll wait for a bit and then see what's up once it errors out
<KarlJoad>
madmaurice: It was building before (yesterday morning (CST), but the white-space change yesterday cahnged something) , but I'm using: `nix-build -E '(import /home/karl/Repos/nixpkgs {}).octavePackages.fits'` to test a package that needs a `patchPhase`.
<V>
I patched my nixos-rebuild to run put sudo in the correct places so i didn't have to run the whole thing as root
<quinn>
abathur++ thanks mate
<{^_^}>
abathur's karma got increased to 17
<pinpox>
colemickens: I set the bios to "UEFI only", so it should be efi anyway right?
<colemickens>
Is there a way to only reinstall the bootloader? Maybe not, since it is Nix after all.
<KarlJoad>
madmaurice: I'm double checking now, but I think quaternion should also have that issue.
heywoodlh has joined #nixos
<colemickens>
You might want to just re-mount everything under /mnt and then just do nixos-install again, tbh.
<{^_^}>
[nixpkgs] @LeSuisse opened pull request #110219 → [20.09] opensmtpd: patches for CVE-2020-35679 and CVE-2020-35680 → https://git.io/JtmMH
<pinpox>
colemickens: Yeah, but it's complicated because my config uses flakes and I coudn't get the installer to pick it up correctly
<KarlJoad>
madmaurice: I had an out-of-date branch up there. You'll need to re-fetch (I had to do a rebase last night for the white-space change there was).
<colemickens>
pinpox: if you can spare the memory, you could only mount /boot, then run install. It will needlessly copy /nix over as well, but then you don't have to mess with mounting root.
<colemickens>
pinpox: the nice thing is, you can do this: nixos-install --system /nix/store/.... and give it your existing system generation
<madmaurice>
@KarlJoad, master?
<colemickens>
pinpox: so you don't even need to really evaluate your config again as long as you know what $(readlink -f /run/current-system) points to
<KarlJoad>
madmaurice: Still octave/fix-broken-packages. The rebase yesterday forced me to do a force-push.
<pinpox>
colemickens: can I do that form the running system or do I need the liveboot?
mkaito has quit [Quit: WeeChat 3.0]
<colemickens>
pinpox: not sure I totally understand, but if things are mounted correctly it shouldn't matter.
<pinpox>
I just checked, /boot is mounted actually. Shuoldn't that then have worked just running the rebuild with --install-bootloader?
<colemickens>
the net result is the same, if you re-run `nixos-install`, it will, under /mnt, setup (/mnt)/nix/store and then setup (/mnt)/boot.
kalbasit_ has quit [Ping timeout: 240 seconds]
<pinpox>
oh, so mount /boot to /mnt/boot then?
<colemickens>
You can skip mounting /nix/store and it will just write into memory (if its' the live disk).
<colemickens>
pinpox: probably
kalbasit has joined #nixos
<colemickens>
pinpox: I'd look for any output that says efibootmgr
<colemickens>
pinpox: the /mnt prefix is for using with nixos-installl, just to be clear
<colemickens>
pinpox: if you're using the env var and nixos-rebuild then your expectations are right, as far as I understand.
<pinpox>
Sorry, I'm a bit confused. This is what I did: booted into my existing installation with the help of the live-usb's bootloader. /dev/sdb1 (the efi partition) is mounted in /boot. I ran nixos-rebuild switch --install-bootloader without errors. But I can't find nix in the bios or boot it spamming f11 to get into the boot-menu (only windows and
<pinpox>
live-stick options available)
<colemickens>
oh I keep forgetting that you're actually booting INTO the config via refind
<colemickens>
that nifty neat refind, I need to play with it soon
<colemickens>
pinpox: okay :)
<pinpox>
yes, sorry if I was confusing you. I did that to make it easier (work in my system as opposed to the installer)
<colemickens>
pinpox: do you have ` boot.loader.efi.canTouchEfiVariables` set?
<colemickens>
pinpox: since you're NOT in the install media, you might not actually be able to update the efi entries
<colemickens>
(this is why I was mentioning looking for some `efibootmgr` output during that `NIXO... nixos-rebuild` command.)
<KarlJoad>
There was no real reason to use `patchPhase`. I just felt that those `sed` changes make more sense as "patches" rather than something to do during build.
<pinpox>
Thanks for the help though
<colemickens>
pinpox: what you're doing makes sense to me and seems straight forward
<pinpox>
Can I maybe specify the bootloader's location ?
<colemickens>
your config looks good from me eyeballing it quickly
<pinpox>
like, maybe since i used the refind it's confused as where to install the bootloader
<colemickens>
I could take a peek at the output from trying to reinstall the bootloader again
<madmaurice>
KarlJoad, ah preBuild might not work either because it does not inherit preBuild.
<pinpox>
ok, let me run nixos-rebuild
<colemickens>
pinpox: maybe, but I doubt it? I'd expect efibootmgr to whine somewhere if it wasn't able to do its job
<{^_^}>
[nixos-homepage] @garbas pushed 14 commits to blog: https://git.io/JtmyU
<colemickens>
pinpox: I guess there is one last big hammer you could use.
<KarlJoad>
madmaurice: Yeah, the buildPhase (pre/postBuild) shouldn't need any major changes too.
<colemickens>
pinpox: it copies the bootloader binary to the "default" uefi location
<colemickens>
but this could also potentially interfere with another bootloader if something else was in that location.
<KarlJoad>
madmaurice: So, I felt patchPhase makes more sense from a "making changes to the source code". It's weird because that build was working before I did the rebase last night and broke afterwards.
<madmaurice>
KarlJoad, question is why is patchPhase not inherited into the attrset for mkDerivation? I can't seem to find the buildOctavePackage function on NixOS/nixpkgs, did you write that?
hnOsmium0001 has joined #nixos
mahogany has joined #nixos
<colemickens>
pinpox: hmph, no new thoughts from me, sorry.
kini has joined #nixos
<colemickens>
maybe I should keep a grub system around to build up some knowledge
<mahogany>
hi everyone, I'm trying to build an overlay which provides an ant/java based program. the ant build itself works and ant tells me "BUILD SUCCESSFUL" but nix says it was unable to produce the output path
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « pappl: 1.0.0 -> 1.0.1 »: https://git.io/JtmSq
<KarlJoad>
madmaurice: The intended effect is that if a package writer doesn't define a new patchPhase, then the stdenv.patchPhase is used. (ltfat uses `patches`). But if they do, then their definition takes precedence. I was unsure about how to do it. A previous version did do it that way.
<KarlJoad>
madmaurice: When using nixpkgs-review on the branch, I get `6 packages failed to build: fits nurbs octclip octproj quaternion strings`, all of which use sed patchPhases. But not ltfat, which uses `patches = [ ./file ]`
<{^_^}>
[nixpkgs] @Ekleog merged pull request #110219 → [20.09] opensmtpd: patches for CVE-2020-35679 and CVE-2020-35680 → https://git.io/JtmMH
<{^_^}>
[nixpkgs] @Ekleog pushed commit from @LeSuisse to release-20.09 « opensmtpd: patches for CVE-2020-35679 and CVE-2020-35680 (#110219) »: https://git.io/Jtm9U
<madmaurice>
KarlJoad, weird. why. does the trace give any information?
codygman__ has quit [Ping timeout: 256 seconds]
<KarlJoad>
madmaurice: error: while eval-ing attr 'buildInputs' of derivation 'shell' at trivial-builders.nix\n 'out.outPath' for customisation.nix\n 'buildInputs' for build-octave-pkg.nix\n 'chooseDevOutputs' lib/attrsets.nix\n infinite recursion in build-octave-pkg line 80.
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #107924 → Fix duti on macOS Big Sur → https://git.io/JLHT0
<{^_^}>
[nixpkgs] @matthewbauer pushed 4 commits to master: https://git.io/JtmQL
<KarlJoad>
Line 80 is the line about `buildInputs`.
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
<V>
imagine you were building the package on one machine to run on another machine. some of the packages you don't need in order to be able to run it, but you do to build it
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/Jtm5v
<KarlJoad>
So, from how I understand it (not in the case of cross-compilation): native = exclusively build-time. build = runtime libraries. I use propagated for binaries needed during runtime.
<pinpox>
colemickens: I fixed my system. Got another hint in #krebs just wanted to let you know in case somone comes across this issue aswell
<{^_^}>
[nixos-homepage] @garbas pushed 88 commits to fix-659: https://git.io/Jtm5k
<colemickens>
What's #krebs and what was the fix pinpox or is it in scrollback?
<pinpox>
the solution was to rm -rf /boot/* and then run nixos-rebuid with --install-bootloader. Config unchanged (touchEfivar = true, installasremvable = false)
<pinpox>
colemickens: ^
nickywhite[m] has left #nixos ["User left"]
<pinpox>
#krebs is a channel here on freenode, originally for krops.nix
<pinpox>
No idea why I had to rm everything in /boot, but that did fix the problem 🤷
<colemickens>
I don't have any idea how that could possibly fix things
<colemickens>
but ok
rb2k has joined #nixos
<KarlJoad>
V: In that case, where should pkg-config go? native- or just build-Inputs?
<V>
KarlJoad: you don't use pkg-config at runtime
<V>
its sole purpose is to provide the compiler with the correct flags and paths
<supersandro2000>
before anyone asks: yes I checked all those PRs by hand
<V>
define "checking by hand"
<KarlJoad>
That's what I thought too. I was told to make pkg-config a buildInput, but I wanted to leave it as a nativeBuildInput for that reason. Is that reasonable?
<pinpox>
colemickens: No idea either, maybe some caching issue?
<supersandro2000>
going to the file and checking if the version there is equal or newer before closing 10 PRs
<V>
KarlJoad: it goes in nativeBuildInputs only
<V>
who told you to put it in buildInputs
<KarlJoad>
I think I was reading the comment wrong on the PR. The reviewer highlighted multiple lines, and I read the wrong one. They were referring to my use of nettle as a nativeBuildInput.
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/Jtm5D
miguel_clean has joined #nixos
<aranea>
supersandro2000: I'd expect more review of non-maintainer package bumps than literally just building them automatically and manually checking that the indeed increase the version
<miguel_clean>
eventhough I have the firewall up iptables hast this first enty in the nixos-fw Chain. is this expected?
<miguel_clean>
nixos-fw-accept all -- anywhere anywhere
<supersandro2000>
aranea: the ones I merged in the last hour had all LGTM from maintainers
<{^_^}>
[nixos-search] @turboMaCk opened pull request #267 → Fix mobile only interactions → https://git.io/Jtm5F
<miguel_clean>
ok sorry forget it, this seems to be for the lo interface only
<V>
https://github.com/NixOS/nixpkgs/pull/110158 this was seen by 1/2 maintainers; depending on the package in question, the maintainers may wish for more than one of them to have reviewed a change
<V>
supersandro2000: I just checked through all the merges you've done, and you merged over half of the bot-made PRs *without a review from the maintainers*
<sorear>
ever, or a specific timespan?
<V>
this evening
fuiltilt has joined #nixos
* sorear
also wants to know what's going on
<V>
what's even the point of having a set of maintainers for a package if someone's just going to come along and semi-automatedly merge the package bump made by a bot with the only conditions being that a) it builds and b) any random person stamped their name on it
<V>
sorear: the pattern almost certainly extends back
zebrag has quit [Quit: Konversation terminated!]
zebrag has joined #nixos
<adisbladis>
V: OTOH it's very common for maintainers to ignore PRs
<V>
I was curious earlier, poked at github a bit and discovered that (at some point earlier today) supersandro2000 was responsible for 403/1000 of the most recent r-ryantm merges
<sorear>
if some subset of maintainers are failing their responsibilities, that sounds like an argument for reassigning packages, not getting rid of the maintainership system
<adisbladis>
sorear: Absolutely.
<adisbladis>
I think our maintainership role is not well defined enough
<V>
additionally, going by merge timestamps, there were like... 85/403 done within a minute of each other? 55/403 done within 30 seconds. 31 done within 20 seconds. 16 done within 15 seconds
<V>
that does not scream "I have actually looked at this properly" to me
<supersandro2000>
most often I do batches
Rusty1 has joined #nixos
<supersandro2000>
I you take a look at new packages merged you will find a similar pattern
* DigitalKiwi
thinks more people should add themselves as a maintainer
<adisbladis>
DigitalKiwi: That doesn't necessarily help if there are no expectations on the maintainer
<DigitalKiwi>
i have a queue right now of a few packages i need to make PRs for that don't have maintainer or i can't find them/they haven't been around :(
<V>
supersandro2000: just verifying that something builds isn't enough
<supersandro2000>
also that discussion already went down like two days ago. With the same result as right now.
<supersandro2000>
so instead of repeating everything: I get it.
<supersandro2000>
another random thought to move the discussion back to something useful: What to do with maintainers that deleted their github accounts?
<supersandro2000>
and have no other mean to contact?
<V>
supersandro2000: but if you get it, why didn't you change anything you're doing?
<V>
understanding a thing and then proceeding to do absolutely nothing with this understanding is about as useless as not understanding it at all
<energizer>
v: what else needs to be verified, in addition to "it builds"?
<DigitalKiwi>
adisbladis: well i figure the more eyes the better
<energizer>
also, take it easy man
<supersandro2000>
V: if you have such strong feeling about any PR go ahead and revert it
<aranea>
energizer: at the very least, you need to read the changelog and watch out for breaking changes
<DigitalKiwi>
energizer: well sometimes they build but then immediately segfault :D
<DigitalKiwi>
or break something :(
<adisbladis>
DigitalKiwi: I'm not disagreeing with you at all :)
<V>
energizer: it's kiiiind of impolite to go merge a bot-generated PR without the maintainers getting to look at something
<adisbladis>
We need both things to happen
<ryantm>
One thing I feel as a Nixpkgs committer is a responsibility toward whatever I merge, so if I merge something and it breaks something, I'll work hard to fix it later. If supersandro2000 feels the same way, I'm not too worried about this specific methodolgy.
<aranea>
if a package has a dedicated maintainer than they're best equipped to handle all this; that's the whole point why we have maintainers
<aranea>
*then
<DigitalKiwi>
znc-playback stopped working for me because it got switched to a different fork...
<sorear>
if a maintainer is unreachable then the package is in an orphaned state until someone else takes ove3r
<V>
currently, if we went and deleted meta.maintainers treewide it would have approximately the same amount of value as it does currently, if we're going to merge things like this
<ryantm>
Merging something doesn't magically make it unavailable to maintainers.
<aranea>
Trying not to break things in the first place is still very much preferable though
<supersandro2000>
if there is something broken with a package a PR I merge send me a message and I'll fix it
<V>
shoot first, ask questions later is not an attitude I want to be applied to a system I am running on all my machines, tyvm
rmcgibbo has joined #nixos
<supersandro2000>
DigitalKiwi: the last change to that package is 4 years old and was a treewide change from fetchgit to fetchFromGitHub
<energizer>
tbh i dont know why breaking master is something that can happen at all
<DigitalKiwi>
V well it is nixos-unstable afterall ;)
<sorear>
DigitalKiwi: who reviews all of supersandro2000's PRs before they get merged to nixos-stable?
<energizer>
nixos-unstable is supposed to mean "tests passed", not "tests didn't run or failed and merged anyway"
<V>
DigitalKiwi: prior to like 3 months ago, I could reasonably use nixos-unstable and be comfortable that the majority of the time, commits that were ending up on there were not going to break my system
<V>
most likely
<V>
as of right now, I'm evaluating which of guix/arch/gentoo I'll hate the least
<ryantm>
I think you are greatly exaggerating past quality.
<V>
no, I'm not
<ashkitten>
supersandro2000: are you verifying that built programs run at least?
vidbina_ has quit [Ping timeout: 264 seconds]
<V>
presumably
<V>
I would expect that at the very least
<ryantm>
I feel like there have been at least about 3 or so major breaking changes each year on unstable. The major one that comes to mind lately is all fonts being unavailable.
<energizer>
why is that a human's job at all?
amk has joined #nixos
erasmas has quit [Quit: leaving]
<V>
nah, the most recent major one was everyone's firefox extensions getting wiped out
<amk>
Howdy, Is there a good way to restrict the paths a service can access? Do nixos containers support it?
<energizer>
what's wrong with "just don't allow breaking changes into master"
<ryantm>
I avoided that one because of zimbatm's tweet :)
<V>
I avoided that b/c I was being lazy with upgrades
<V>
and then someone who had the misfortune of having it happen to her poked me about it
<rmcgibbo>
supersandro2000: Is your normal workflow for running `nixpkgs-review-checks` that you run nixpkgs-review pr <number>, and then in the new shell you run `nixpkgs-review-checks` followed by post-result?
<samueldr>
ryantm: the fonts thing was a bit more nuanced than "no fonts on unstable lol", it was that it could poison the well of fonts when running things from mixed origins :)
<ryantm>
V: Because I want to consider level of involvement in the project when considering how much to value your opinion on how the project operates.
<supersandro2000>
rmcgibbo: it is written to PROMPT_COMMAND which runs it automatically
<V>
ryantm: I see how it is
<DigitalKiwi>
ryantm: right now i think plasma5 is broken :(
<edef>
ryantm: if we give V a commit bit and she starts mass-merging stuff, you'll take her seriously?
<gchristensen>
I agree with V
<supersandro2000>
when I first tried NixOS plasma5 was broken, too and you could not change the mousepad settings at all.
<edef>
because that can be arranged for
<edef>
happy to give a demonstration of Goodhart's law
<ryantm>
edef: Nope
<edef>
you're welcome to demean me as well
<supersandro2000>
I already said that I took that feedback to my heart and stopped doing it.
<colemickens>
(tbh that plasma thing just sounds like plasma)
<rmcgibbo>
supersandro2000: Okay. I'm still learning all the shell-fu. If you use `nixpkgs-review` with `--no-shell` and `--post-result`, there's no opportunity for nixpkgs-review-checks to intercept the report between the time at which it's created and the time at which it's uploaded, correct?
<supersandro2000>
rmcgibbo: yeah but my script would not work with that setting and I am not doing this.
<edef>
but if we're bringing the community engagement down to that level, i don't think my negative energy will be directed towards the people doing thoughtful contributions in smaller number
<ryantm>
edef: Demean?
<rmcgibbo>
supersandro2000: Totally understood. I'm just trying to figure out how it works. Thanks!
<supersandro2000>
If anyone has constructive feedback send me a DM and I read it later.
<DigitalKiwi>
supersandro2000: well it was working but now i can't get xmonad+plasma5 to work it doesn't have plasmashell and it tries to start kwin which then coredumps :(
<DigitalKiwi>
i
<edef>
ryantm: idk, i'd like to know your standards for what counts as enough contribution to have a valid opinion
<DigitalKiwi>
i'm getting rid of anything that might have a different qt version and cleaning cache and config to see if that helps
<ryantm>
edef: It's not something specific, just a feeling of trust that develops over years of seeing people around. Like I know you've been around the project a long time, so I trust you.
<ryantm>
edef: It can of course go the other way, if I see them around but doing dumb stuff :)
<edef>
yes, but i think that we have non-insignificant turnover in contributors and relatively little new blood
<DigitalKiwi>
i've been trying to git bisect it but that ended up showing me a commit that couldn't have done it :(
<gchristensen>
to the various people sending me PM right now, a lot of people are PMing me things right now, and my brain is somewhat overloading, and it might be good to be more upfront in the moment
<gchristensen>
(sorry)
tokudan has quit [Remote host closed the connection]
<DigitalKiwi>
colemickens: lol yeah xmonad doesn't even quite work right (i have to change the menus to the "alternative" one because of a bug in Qt that makes windows not appear at the right layer... and text input doesn't work in like krunner all of the time >.>