<lukegb>
aleph-: they should charge for the option to remind them to ship the package
<aleph->
Heh
rajivr has joined #nixos-chat
<DigitalKiwi>
they do it's called the checkout button ;p
endformationage has quit [Quit: WeeChat 2.9]
waleee-cl has quit [Quit: Connection closed for inactivity]
supersandro2000 has quit [Killed (kornbluth.freenode.net (Nickname regained by services))]
supersandro2000 has joined #nixos-chat
c4rc4s has joined #nixos-chat
<eyJhb>
gchristensen: Regarding the appstore for modules. I think I have some basic idea of how it could be done. I am however not sure, it is the best. But basically just a github repo, which adds a extra option to nixos options, which lets you specify a list op options from that repo that you want. ANd then doing it kind of like the nixos-community. But as I am writing this, I remember that there was
<eyJhb>
some talk about doing it "web" based, where devs could register and upload stuff... But that should be doable as well I suppose.
cole-h has quit [Ping timeout: 240 seconds]
evanjs has joined #nixos-chat
__monty__ has joined #nixos-chat
abstrn has joined #nixos-chat
zgrep has joined #nixos-chat
luxemboye has joined #nixos-chat
h0m1 has quit [Ping timeout: 245 seconds]
h0m1 has joined #nixos-chat
endformationage has joined #nixos-chat
<hexa->
hm, can I have two systemd units use the same DynamicUser?
<hexa->
like with a slice or something?
<lukegb>
I don't even think you need a slice
<lukegb>
As long as the names match I think they'll end up with the same UID
<hexa->
currently when they share a StateDirectory the permissons get screwed up
<hexa->
-rw------- 1 64779 64779 24K May 3 14:39 pinnwand.db
<lukegb>
Hmm, even with explicit matching names?
<hexa->
-rw------- 1 pinnwand pinnwand 24K May 3 14:38 pinnwand.db
<gchristensen>
so they're not alarms but notifications that my buildkite pipeline finished successfully ... I want this sort of notif for most things but ...not this one :D
<philipp[m]>
I do get very suspicious when things work because that usually just means they are broken in more complicated ways.
<pie_>
its kind of ridiculous that man with nix-shell still doesnt always work
<{^_^}>
nix#4702 (by sternenseemann, 3 weeks ago, open): nix shell: set MANPATH for installables that have a man dir
<gchristensen>
I think mkderivation would need to change for that, we're looking for a mass-rebuild in stdenv, maybe we can fix it
<gchristensen>
yeah, changing nix is totally the wrong approach
<lukegb>
does shellhook get evaluated for -p?
<pie_>
gchristensen: cant we fix it in the interim at least
<pie_>
(which of course means it will never get fixed properly eh?)?)
<gchristensen>
lukegb: some way for nixpkgs to inform nix on what to do here is the right approach, regardless :)
<lukegb>
yeah
<lukegb>
I agree that we probably don't want to start pushing e.g. magic Qt logic into nix proper
<gchristensen>
+1
<pie_>
sterni: here's another thing about manpath dunno if related but the comment about it overriding stuff seems a bit of a pain https://github.com/NixOS/nix/pull/2565
<{^_^}>
nix#2565 (by jvassev, 2 years ago, open): Saner defaults for MANPATH
<pie_>
gchristensen: can you detect shell?
<pie_>
does shell have anything for looking for specific outputs?
<gchristensen>
nix-shell always uses bash
<pie_>
here's another one
<pie_>
also just realized im in -chat again
<pie_>
so i just keep doing this i guess? MANPATH=$(nix-build --no-out-link -E "with import <nixpkgs> {}; binutils.man")/share/man man objdump
<sterni>
lukegb: but we kind of need to, the way installables work this is not fixable
<sterni>
and tbh I haven't gotten the impression that anybody is really interested to fix this in the nix stable cli
<sterni>
apart from hack it into nixpkgs somehow
<sterni>
which I am sceptical about since we'll still loose the feature with the future nix CLI
<sterni>
and we don't need manpath in the build environment
rajivr has quit [Quit: Connection closed for inactivity]
rj has quit [Quit: rj]
rj has joined #nixos-chat
rj has quit [Client Quit]
rj has joined #nixos-chat
<andi->
IMHO we should just stop using nix-shell and write a tool that configures your shell based on a JSON document that you emit from mkShell' (to be invented) in nixpkgs. This would a) remove the bash restrictions and b) allow defining the env vars without touching Nix. It shouldn't be the interpreter that has the knowledge. It should be the SCM'ed source repo that you are working on.
rj has quit [Client Quit]
h0m1 has quit [Quit: WeeChat 3.1]
<hodapp>
another language? >_>
rj has joined #nixos-chat
h0m1 has joined #nixos-chat
<samueldr>
IIRC in "nix next" there is an attempt at something to fix the same issue
<andi->
Why another language? I imagine `mkShell' { inputs = [ pkgs.firefox ]; }` to emit a JSON doc stating `PATH=....firefox/bin/:...`
<samueldr>
so that it's not a really-leaky abstraction over bash
rj has quit [Client Quit]
<samueldr>
not sure if it's been implemented yet or not, as I don't follow the nix development
rj has joined #nixos-chat
<andi->
What is nix next?
<samueldr>
what I call "whatever master is, and the plan for the future"
<andi->
Yet another unsteable experiment that is halfdone and thought out?
<samueldr>
because we don't have a better name
<andi->
oh, ok
<samueldr>
that's a strong wording you're using, hopefully it won't apply :)
<samueldr>
and I won't say "flakes" since I don't know if it was part of flakes or not
<andi->
I am just (still) upset about everything flakes. I am not even trying to hide my opinion of that pile of garbage anymore. There is nothing to loose for me.
<samueldr>
would be interesting to split out the improvements in "nix next" and rebase them onto 2.3 to produce a 2.4
<samueldr>
(with stabilization)
<samueldr>
I literally have no idea what improvements there are
<samueldr>
detached from the flakes experimental stuff
<andi->
Given the state of the "new" cli I don't think that it would be great to make a 2.4 release without flakes (or with it).
<joepie91>
andi-: do you have documentation of your issues with flakes somewhere?
<andi->
see discourse that is the most recent post I made.
<samueldr>
I don't have a clue about the current development, so are you implying all the improvements/changes/fixes in 2.4 are entirely done behind flakes?
<andi->
no, not all of them. I am pretty sure the while refactor of fetchers and goals is good.
<samueldr>
that's mostly it, what changes can we backport into "nix current" and make "nix current++" happen so that "nix next" is less of a total change
<samueldr>
since AFAICT we're not close to a "nix next" release? though I don't know either, maybe it's slated for tomorrow and I'm simply not aware
<andi->
Nobody knows but one.
<samueldr>
(trying to move this into -dev)
<joepie91>
lukegb: (cont. from -dev) I feel like you would have enjoyed TideSDK
<joepie91>
before it was eaten by Electron
<lukegb>
heh
<joepie91>
mixed-language runtime kernel, shared scope between JS, Python, Ruby, and PHP, if I'm remembering correctly
<joepie91>
each of them could access each of the others
<joepie91>
all of it hosted within a webkit thingem
<joepie91>
sort of a multi-language predecessor of Electron, but 200% cursed
<cole-h>
has anybody else been getting multi-second "lag"s on IRC recently?
<samueldr>
probably
<cole-h>
just wondering because internet went out last night, and now I'm seeing multi-second lags being reported by weechat almost every minute
<cole-h>
wondering if it's a coincidence, or related
<andi->
Haven't seen that here in a while
<cole-h>
maybe it's the freenode server I'm connected to. /me reconnects
<hexa->
infinisil: people keep talking in the channel, guess I need a few minutes alone to figure /that/ out
<hodapp>
gchristensen: as much as I've liked it, some things in it feel very batteries-not-included
<hexa->
but it sounds somewhat better
<infinisil>
Nice
<hexa->
sadly mumble 1.4 and pipewire dont seem to get along well
<hexa->
ok, nvm, mumble 1.3.4 and pipewire is also broken now <.<
<gchristensen>
hodapp: is there a function like: join ", " .Params.authors where authors is defined in yamlfront matter with authors:\n - foo\n - bar ?
<gchristensen>
delimit
hax404 has joined #nixos-chat
rj has quit [Quit: rj]
rj has joined #nixos-chat
rj has quit [Client Quit]
supersandro2000 is now known as Guest35175
Guest35175 has quit [Killed (card.freenode.net (Nickname regained by services))]
supersandro2000 has joined #nixos-chat
__monty__ has quit [Quit: leaving]
<elvishjerricco>
Sometimes I feel like ZFS's slog and L2ARC features need big flashing warning signs that say "these probably don't do what you think they do"
<elvishjerricco>
People on Reddit always seem to get it wrong
<hexa->
can you expand on that for a casual user?
<samueldr>
slog is what you'll do when you use L2ARC ?
<samueldr>
>> a spell of difficult, tiring work or travelling.
<samueldr>
"it would be a hard slog back to the camp"