worldofpeace_ changed the topic of #nixos-dev to: #nixos-dev NixOS Development (#nixos for questions) | NixOS stable: 20.03 ✨ https://discourse.nixos.org/t/nixos-20-03-release/6785 | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | https://r13y.com | 19.09 RMs: disasm, sphalerite; 20.03: worldofpeace, disasm | https://logs.nix.samueldr.com/nixos-dev
bgamari has quit [Ping timeout: 272 seconds]
alp has quit [Ping timeout: 272 seconds]
bgamari has joined #nixos-dev
<jtojnar> ugh, nixos org does not allow github cli oauth access apparently
<gchristensen> huh?
Scriptkiddi has quit [Quit: killed]
das_j has quit [Quit: killed]
ajs124 has quit [Quit: killed]
ajs124 has joined #nixos-dev
das_j has joined #nixos-dev
Scriptkiddi has joined #nixos-dev
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-dev
drakonis has quit [Quit: WeeChat 2.8]
alp has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
* jtojnar uploaded an image: fractal-pasted-image (95KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/ZCOLhcElGQzfStWRIQYFHboS >
__monty__ has joined #nixos-dev
alp has quit [Ping timeout: 272 seconds]
alp has joined #nixos-dev
Jackneill has quit [Ping timeout: 256 seconds]
disasm has quit [Ping timeout: 260 seconds]
Jackneill has joined #nixos-dev
nschoe has joined #nixos-dev
bgamari has quit [Quit: ZNC 1.7.5 - https://znc.in]
bgamari has joined #nixos-dev
bgamari has quit [Remote host closed the connection]
{`-`} has joined #nixos-dev
bgamari has joined #nixos-dev
<bennofs[m]> hmm, I think a public log of manual actions for hydra would be nice
nschoe has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
nschoe has joined #nixos-dev
nschoe has quit [Ping timeout: 260 seconds]
nschoe_ has joined #nixos-dev
nschoe_ has quit [Client Quit]
nschoe has joined #nixos-dev
disasm has joined #nixos-dev
<xfix> okay, i decided to analyze why LiveCD is so huge, and it seems like netronome firmware is 140MB or so
<xfix> that seems to be a lot for firmware for hardware that is very very very very unlikely to be found on end-user machines
alp has quit [Ping timeout: 244 seconds]
<xfix> but other than that potential savings are non-trivial
bgamari has quit [Quit: ZNC 1.7.5 - https://znc.in]
bgamari has joined #nixos-dev
bgamari has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 272 seconds]
bgamari has joined #nixos-dev
alp has joined #nixos-dev
<emily> wow, does that firmware just contain an entire embedded OS distribution or something
_e has quit [Quit: WeeChat 2.7.1]
_e has joined #nixos-dev
tilpner has quit [Quit: tilpner]
hexa- has quit [Quit: WeeChat 2.7.1]
hexa- has joined #nixos-dev
terrorjack has joined #nixos-dev
_e has quit [Quit: WeeChat 2.7.1]
cole-h has joined #nixos-dev
alp has quit [Ping timeout: 265 seconds]
justanotheruser has quit [Ping timeout: 272 seconds]
justanotheruser has joined #nixos-dev
drakonis has joined #nixos-dev
<xfix> emily: i wouldn't be too surprised
<xfix> of course, it's different models and all that sum up to 140MB
alp has joined #nixos-dev
<gchristensen> cole-h: I just want to call out how much I appreciate your involvement, interest, and care lately. It has made a difference to the project and me personally, thank you for it. Things have felt a bit less "alone" in some ways -- and it is nice to have help on my "not me, team" project
<cole-h> :D
<LnL> cole-h++
<{^_^}> cole-h's karma got increased to 43.00000000000000004
<cole-h> <3 gchristensen Happy to help. I've always wanted to be involved in larger open-source projects, so I feel like I've "graduated" in a way, now :^)
<{^_^}> gchristensen's karma got increased to 287
<pie_> wait what <{^_^}> cole-h's karma got increased to 43.00000000000000004
<cole-h> infinisil: Added some spice to the karma messages for "rare" occasions
<pie_> ;p
<infinisil> I swear I didn't! It's these damn floats again!
<gchristensen> lmao
<adisbladis> Where did cole-h come from? Recently
<adisbladis> I've never head of this person, now produces awesomeness all over the place
<cole-h> adisbladis: lol :D gchristensen was soliciting Rust help here a few months ago I think
<cole-h> Something about making errors better (because we use/d a whole lot of expects and unwraps which are runtime panics)
<cole-h> (Which I still haven't done lol)
<gchristensen> +1
<cole-h> I started by picking up LnL's old Rust 2018 PR (because 2015 edition, while it wasn't bad, is annoying to write), making opinionated changes, and then it kinda spiraled from there
ixxie has joined #nixos-dev
<pie_> is there a reason we dont have a current-stable channel?
<pie_> or metadata somewhere that points to what the current release is?
orivej has joined #nixos-dev
<adisbladis> pie_: !!
<adisbladis> I need this
<adisbladis> I have software I consider supported on unstable + "current stable"
<adisbladis> But have to manually bump branches on CI to follow what stable is
<pie_> adisbladis: im not volunteering to implement it :(
<pie_> but i would like this
<pie_> currently cole-h has suggested the following; 'xD
<pie_> <cole-h> `curl 'https://raw.githubusercontent.com/NixOS/nixos-homepage/master/Makefile' | head -n1` => `NIXOS_SERIES = 20.03`
<cole-h> Don't forget the `| cut -d' ' -f 3` to get the version alone :D
<pie_> If we had something like https://raw.githubusercontent.com/NixOS/nixpkgs-channels/nixos-unstable/.version on unstable that "points" in some shape or form to the current release i think that would be viable but maybe its easy to just have a current-stable channel or something?
<{^_^}> #87124 (by deliciouslytyped, 8 seconds ago, open): Add a nixos-stable channel
<pie_> TBH I could even argue for something like it being the default, but because of possible breakages across releases, it's probably not reasonable overall.
<abathur> the other day I re-ran an old CI job that succeeded 2 years prior and it failed; I've had a back-of-mind curiosity since about if/how others occasionally run CI against moving targets like channels/branches
<cole-h> I use niv to pin stuff (:
<cole-h> I imagine that will be solved (or at least alleviated) by flakes
<abathur> well, right, but it'd still be nice to know when breaks are popping up
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-dev
<cole-h> How do you mean breaks?
<abathur> any new dep/build/test failure at a project commit that previously succeeded but starts failing as nixpkgs rolls on into the future?
<cole-h> Got it.
<cole-h> Running CI for your previously-run CI runs?
<abathur> heheh
* julm didn't know about https://nix-community.org/ nice!
<abathur> well, I was revisiting an early phase project I haven't had time to touch for 2 years, and the tests weren't running--but they ran clean the last time I ran CI
<abathur> so I was re-running it as a sanity check
<abathur> and have had a little percolating thought since that it'd be nice to have a feedback drip of some sort there
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-dev
<abathur> I think at least in the case of travis-ci I could set up a branch with recurring cron jobs to handle the Nixpkgs fraction of that, but sometimes I'm also building an explicit version/ref of a dependency; might be nice to let them drift and get feedback
<abathur> I guess it can also be a little bit of both if I'm specifying a rev in an override and in theory package behavior could change under it
* julm just saw that there is a Mumble for talking about Nix: mumble://mumble.nix-community.org
<drakonis> ah how nice, when did nix-community gain a website?
<drakonis> late last year, nice.
<julm> gchristensen: in case you haven't seen yet: https://elis.nu/blog/2020/05/nixos-tmpfs-as-root/
<gchristensen> yep, it is great!
alp has quit [Ping timeout: 246 seconds]
alp has joined #nixos-dev
ixxie has quit [Ping timeout: 246 seconds]
<misuzu> i did something like that using NIXOS_LUSTRATE code
<pie_> adisbladis: your comment would help
<{^_^}> #87124 (by deliciouslytyped, 2 hours ago, open): Add a nixos-stable channel
<adisbladis> pie_: ✔
<infinisil> Regarding the discussion on getting statistics on packages in #nixos-chat, I have the following idea
<infinisil> I have a patch to nixpkgs that allows derivation expressions to know which nixpkgs attribute path they came from
<infinisil> So pkgs.haskellPackages.xmonad.__nixpkgsAttrPath would be [ "haskellPackages" "xmonad" ]
<infinisil> Now we could make use of this information by changing Nix itself to (when opted-in), automatically upload stats on these attribute paths of all built derivations
<infinisil> s/built/realized
<clever> infinisil: nix-index works by just re-evaling the whole <nixpkgs> and then checking that against a list of storepaths
<clever> infinisil: to map a pack backwards to an attr
<infinisil> Yeah but that won't help here I'm pretty sure
<infinisil> With Nix submitting these attribute paths, we would only expose what *nixpkgs* packages people use, not internal company ones
<infinisil> And it wouldn't include which store path exactly, only the attribute path, without any version
<infinisil> gchristensen: sphalerite: ^ Thoughts on this?
<arianvp> I've re-read this section 5 times and i still dont understand
<arianvp> 1. am I supposed to not use buildInputs and nativeBUildInputs anymore?
<sphalerite> arianvp: if you want it on the PATH, use nativeBuildInputs
<arianvp> but it says I should use depsBuildHost instead
<sphalerite> arianvp: alternatively: enable strictDeps and see if it breaks. If not, you probably did it right. :p
<arianvp> it seems to suggest buildInputs and nativeBuildInputs are deprecated
ehmry has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<arianvp> > Nixpkgs is now structured so that each depsFooBar is automatically taken from pkgsFooBar
<{^_^}> undefined variable 'Nixpkgs' at (string):309:1
<arianvp> this part I dont understand
<arianvp> if I have
<sphalerite> "In short, each list of dependencies for "host → target" of "foo → bar" is called depsFooBar, with exceptions for backwards compatibility that depsBuildHost is instead called nativeBuildInputs and depsHostTarget is instead called buildInputs."
<arianvp> nativeBuildsInputs = [ pkgs.foo ]; is this somehow magically rewritten to pkgs.pkgsBuildHost.foo ?
<arianvp> how?
<sphalerite> so buildInputs and nativeBuildInputs is right
<sphalerite> that bit I don't understand :)
<arianvp> that sounds totally like black magic to me
<clever> i'm also interested in how nativeBuildsInputs magically rewrites
<arianvp> and i dont understand
<arianvp> im very happy this section exists now!! used to be all undocumented. but it's still very difficult and im not sure if it's without errors
<clever> infinisil: .nativeDrv and .__spliced dont exist
<clever> > builtins.attrNames pkgsCross.raspberryPi.hello
<{^_^}> [ "__ignoreNulls" "all" "args" "buildInputs" "builder" "cmakeFlags" "configureFlags" "depsBuildBuild" "depsBuildBuildPropagated" "depsBuildTarget" "depsBuildTargetPropagated" "depsHostHost" "depsHostH...
<arianvp> > builtins.attrNames pkgs.hello
<{^_^}> [ "__ignoreNulls" "all" "args" "buildInputs" "builder" "configureFlags" "depsBuildBuild" "depsBuildBuildPropagated" "depsBuildTarget" "depsBuildTargetPropagated" "depsHostHost" "depsHostHostPropagated...
<clever> infinisil: the problem, is that i'm finding some exprs that do `RUSTC = rustc;` which gets the target rust, not the host rust
<clever> infinisil: and if i change the package to accept buildPackages via callPackage, it heavily complicates .override { rustc = ...; }
<clever> so now what?
<arianvp> ugh why this all so complicated :')
<arianvp> > pkgs.hello.__spliced
<{^_^}> attribute '__spliced' missing, at (string):309:1
<infinisil> clever: Not entirely sure where __spliced appears, but it's in there somewhere I'm sure lol
<arianvp> just seems callPackage should be cross-aware instead
<arianvp> this thing seems like a bit of a hack after the fact to not break all the callPackage things :P
ehmry has joined #nixos-dev
<arianvp> is there an easy way to (ab-)use patchShebangs to patch other things than shebangs?
<arianvp> in my case I want to patch paths in systemd ExecStart lines
<clever> arianvp: substituteInPlace and friends
<cole-h> Or even sed
<arianvp> clever: with substituteInPlace I need to know the name of the binary beforehand; which I dont :)
<cole-h> Why don't you know the binary beforehand? Seems like a strange limitation...
<cole-h> You might be interested in https://github.com/abathur/resholved
ixxie has joined #nixos-dev
<arianvp> because you dont either with patchShebangs. you discover the binary name; and then you look up in PATH for the relevant nix store path
<pie_> adisbladis: cole-h: thanks for commenting
__monty__ has quit [Quit: leaving]
<cole-h> arianvp: Fair enough. I'd be interested to see if you turn up anything, but I don't have any ideas (sorry).
ixxie has quit [Ping timeout: 246 seconds]
julm has quit [Quit: Lost terminal]
julm has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
justanotheruser has quit [Ping timeout: 256 seconds]
justanotheruser has joined #nixos-dev
ixxie has joined #nixos-dev
alp has quit [Ping timeout: 272 seconds]
ixxie has quit [Quit: Lost terminal]