gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
iqubic has quit [Remote host closed the connection]
iqubic has joined #nixos-chat
iqubic has left #nixos-chat [#nixos-chat]
sir_guy_carleton has joined #nixos-chat
__monty__ has quit [Quit: leaving]
drakonis has quit [Read error: Connection reset by peer]
<gchristensen> nonullpointers.com/posts/2019-01-07-a-year-of-nixos.html
<samueldr> not many operating systems or distros where the users are pointing out "things broke, but eh, things got better" most of the time it's "things broke and I hate things"
iqubic has joined #nixos-chat
<jackdk> the dark side of "it just works" is "it just breaks"
<dmc> lol
<samueldr> it reproducibly works!
<samueldr> it reproducibly breaks too
<gchristensen> haha yes
iqubic has quit [Remote host closed the connection]
drakonis has joined #nixos-chat
jasongrossman has quit [Remote host closed the connection]
sir_guy_carleton has quit [Quit: WeeChat 2.2]
<infinisil> Since my last nixpkgs update when I watch a video in firefox, it doesn't prevent the screen from going black anymore (after 10min or whatever)
<infinisil> If I had the motivation I could totally bisect the issue :)
<srhb> infinisil: It's a feature :|
<infinisil> huh?
<gchristensen> can confirm, thats a feture
<gchristensen> ff is supposed to prohibit dpms when multimedia is playing
<infinisil> But.. it's not doing that as of now?
<gchristensen> oh... doesn't.
<gchristensen> reading comprehension.
<srhb> My reading comprehension failures are contagious, apparently...
<infinisil> Haha, phew
<infinisil> So yeah, the screen goes black for me after ten minutes of watching media in firefox
<infinisil> I guess I should test at least if it's a firefox issue or an X issue
<gchristensen> https://github.com/ludios/ityntolm/commit/3269195af2857e78aa4642eacf6fb6e2c9e387c5 <- these are cool hooks to attach to the screen locker... but not sure about the allow-all one, probably nicer to remember the disabled ones and only re-enable those
<srhb> You should also be able to use dbus-monitor to see if firefox is trying to prevent it.
<infinisil> Oh and I guess I could set the dpms time to something low to test
drakonis has quit [Read error: Connection reset by peer]
<elvishjerricco> infinisil: I realized a reason for the fixed UIDs/GIDs in NixOS. `nesting.children`. You can boot completely different configurations on the same system, so if you use the same root, you'd want the same IDs
<infinisil> Doesn't /var/lib/nixos/{g,u}id-map already make sure for that to be the case?
<infinisil> elvishjerricco: ^
<elvishjerricco> infinisil: Good point, yea probably.
<gchristensen> not cross-machine
<gchristensen> if you use certain databases, you need every system to have the same uid/gids for that service
<ldlework> Anyone using Pomodoro?
<ldlework> Tried it for the first time today.
<gchristensen> did it work?
<ldlework> Yeah
<ldlework> It was extremely hard
<ldlework> I wanted to switch back to IRC almost continuously
<ldlework> lol
<gchristensen> haha
<ldlework> But I fixed a huge bug in an emacs package of mine :D
<gchristensen> oh hot dang, nice!
<ldlework> Well more of an optimization
<infinisil> gchristensen: Well if you have such requirements you should really be using some service to coordinate the machines, and that should also make sure the permissions are correct. It feels wrong for NixOS to have to handle that case
<gchristensen> to have NixOS avoid impurities..?
<infinisil> We don't have unlimited ids, services change, and systemd has great support for temporary ids. It feels wrong to have to make ids static.
<gchristensen> yeah. I think for most things, dynamic IDs are fine
<infinisil> If applications really depend on what id number they have then they should fix that
<infinisil> If you can show me some distributed database or whatever that points out in their docs that they expect all machines have the same ids, that would be a good reason
<infinisil> And I'd gladly keep static uid/gid support for just those
<gchristensen> seems reasonable
<infinisil> Cool, I'll probably add that to my document at some point
<gchristensen> and I suppose it is possible to have users pin them on their own
<gchristensen> for other annoying cases
<infinisil> Good point
<gchristensen> ldlework: https://github.com/NixOS/nix-mode/pull/71 does this look reasonable?
<{^_^}> nix-mode#71 (by grahamc, 12 seconds ago, open): Prioritize closing } over regular indentation
sir_guy_carleton has joined #nixos-chat
lassulus_ has joined #nixos-chat
lassulus has quit [Ping timeout: 246 seconds]
lassulus_ is now known as lassulus
iqubic has joined #nixos-chat
lopsided98 has quit [Quit: Disconnected]
lopsided98 has joined #nixos-chat
endformationage has quit [Quit: WeeChat 2.3]
<jackdk> I'm looking at https://github.com/NixOS/nixpkgs/issues/53626 - scotty appears to depend on hpc-coveralls, which has too-tight bounds for ghc 8.6 and recent aeson. scotty doesn't actually need the package unless you set a flag. Is there a way to hide the dependency from hackage-packages.nix and not build with the flag?
<{^_^}> #53626 (by endgame, 8 minutes ago, open): haskell.packages.ghc863.hpc-coveralls fails to configure (deps too restrictive)
sir_guy_carleton has quit [Quit: WeeChat 2.2]
jackdk has quit [Ping timeout: 246 seconds]
pie_ has joined #nixos-chat
<elvishjerricco> So, Haskell derivations don't parallelize very well at all. I've never seen them benefit from more than `--cores 4`. Is there anything I can do so that I don't have to add `--cores 4` every time I work with Haskell derivations?
jasongrossman has joined #nixos-chat
__Sander__ has joined #nixos-chat
<ivan> my per-machine wireguard config generator (incl picking the right endpoint) https://gist.github.com/ivan/05845e04a469f081f7cea55e28b7180e
<ivan> this replaced a lot of Elixir code I had
iqubic has quit [Ping timeout: 260 seconds]
<ivan> I figured out how to set the xfce4-panel background color :-) https://gist.github.com/ivan/82e525ab29f5a653511ec0bcdf2fd378
<ivan> my IME toggle is wired up to set the color
<dtz> elvishjerricco: I don't know but it hurts pretty bad on big multicore machines D:
<dtz> haha yeah man it's crazy :(
<gchristensen> I remember the first time I could run «make -j4» --- it was so exciting. hard to believe we're up to make -j96
<dtz> for real
<infinisil> gchristensen: *you*'re up to -j96, I'm still here with my lousy 8 cores!
<gchristensen> I am as personally up to -j96 as you
<gchristensen> it isn't anything I can use for myselsf
<cransom> if your io is really bad, you can still -j96 and it would help.
<gchristensen> haha
<cransom> (we used to -j4 on single core machines because the disks were pokey enough that it did make things faster)
hedning has quit [Remote host closed the connection]
endformationage has joined #nixos-chat
<etu> gchristensen: Are you around?
<gchristensen> yeah
<etu> gchristensen: I think I've found another regression: https://ptpb.pw/-XRN/nix
<gchristensen> 3-7
<etu> So the first one, ok, that's the most common one I've seen :)
<etu> I should make a testcase of it
<gchristensen> :)
<gchristensen> I played a few minutes last night on issue #43 but it was too much for me
<{^_^}> https://github.com/NixOS/nixpkgs/pull/43 (by cpages, 6 years ago, merged): kobodeluxe: fix compilation errors
<etu> Fooled by the link :D
<etu> gchristensen: yeah, that seems like a tough one
__Sander__ has quit [Quit: Konversation terminated!]
fpletz has joined #nixos-chat
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-chat
<etu> gchristensen: yeah, I found what was up with the strings nix-mode#72 and the solution in nix-mode#73
<{^_^}> https://github.com/NixOS/nix-mode/issues/72 (by etu, 17 minutes ago, open): Incorrect indentation of multi-line strings
<{^_^}> https://github.com/NixOS/nix-mode/pull/73 (by etu, 1 minute ago, open): Check the correct element from syntax-ppss
emily has joined #nixos-chat
drakonis has joined #nixos-chat
<gchristensen> etu: specifically, the lines in that text should have different indentations and nix-mode should not disturb them
<elvishjerricco> gchristensen: Do you think we should limit NIX_BUILD_CORES in Haskell's generic-builder until issues like those are resolved?
<gchristensen> yeah probably
<gchristensen> cc dtz ^
<elvishjerricco> gchristensen: I'll open an issue in GitHub for discussion
<gchristensen> great
<gchristensen> ping peti :)
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-chat
<etu> gchristensen: ah, cotcha
<pie_> current status of my preliminary run at doing wine stuff with nix: https://bpaste.net/show/850228ac7ab7 https://bpaste.net/show/a763a5542b45 . it kind of works but i cant figure out why my overlayfs is immutable still , also yes i know it needs root and i need to find a way around that at some point, maybe by not using overlayfs, idk :P
<pie_> err or rather maybe its not immutable but it has weird issues with opening and writing to existing files
<infinisil> Youtube down!
<infinisil> And back again
<pie_> must be tapping the tubes
endformationage has quit [Quit: WeeChat 2.3]
endformationage has joined #nixos-chat
<etu> gchristensen: The reason for nix-mode to normalize all the lines of a string to the same level is due to the fallback-identation to put previous lines indetation level on the current line.
<etu> or now i'm confused again :D
<etu> elisp...
<ldlework> gchristensen: i looked at nix-mode but things have changed too much i didn't want to get into it again sorry for not responding
<gchristensen> no worries
<gchristensen> thanks
<elvishjerricco> Hm. Do you not have to set enableParallelBuilding=true; anymore? This always yields `4` for me: `nix-build -E '(with import <nixpkgs> {}; stdenv.mkDerivation { name="hi"; buildCommand = "echo $NIX_BUILD_CORES"; enableParallelBuilding = false; })'`
<elvishjerricco> gchristensen: Huh. Would've expected it to actually change `NIX_BUILD_CORES`
<gchristensen> ah
<gchristensen> NIX_BUILD_CORES is set by Nix, and Nix doesn't actually know / care about enableParallelBuilding
<elvishjerricco> Got it
<fpletz> this way changes to NIX_BUILD_CORES are impure and don't trigger rebuilds on change
<elvishjerricco> right
<elvishjerricco> Anyway, there's no staging for haskell, right? For mass haskell rebuilds, I just need to point the PR at staging?
<gchristensen> right
jackdk has joined #nixos-chat
sir_guy_carleton has joined #nixos-chat
sir_guy_carleton has quit [Ping timeout: 258 seconds]
sir_guy_carleton has joined #nixos-chat