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
<worldofpeace> I just reverted to not using the path I provided in passthru for that option
<{^_^}> #73966 (by d-goldin, 22 weeks ago, merged): make-options-doc: fix string context issues
<cole-h> Well, at least it evals, now...
drakonis has quit [Quit: WeeChat 2.8]
ris has quit [Ping timeout: 246 seconds]
drakonis has joined #nixos-dev
bhipple has joined #nixos-dev
das_j has quit [Quit: killed]
Scriptkiddi has quit [Quit: killed]
ajs124 has quit [Quit: killed]
Scriptkiddi has joined #nixos-dev
ajs124 has joined #nixos-dev
das_j has joined #nixos-dev
orivej has quit [Ping timeout: 240 seconds]
<elvishjerricco> So why does initrd use busybox instead of util-linux and coreutils? Looks to me like it'd only take one or two more megabytes
<samueldr> inertia and baggage!
<samueldr> that would be my guess
<elvishjerricco> samueldr: Would it be a bad idea for me to try changing it?
<samueldr> I don't know if it's a good idea, but it's definitely not a bad idea
<samueldr> could there be something part of busybox not part of those two that is used%
<samueldr> ?*
<elvishjerricco> Maybe... Doubt it though
bhipple has quit [Ping timeout: 246 seconds]
drakonis has quit [Quit: Leaving]
* colemickens casually gestures at #nixos-systemd too
cole-h has quit [Quit: Goodbye]
alp has joined #nixos-dev
pie_[bnc] has quit [Ping timeout: 260 seconds]
page_ is now known as page
pie_[bnc] has joined #nixos-dev
tilpner has quit [Quit: tilpner]
alp has quit [Ping timeout: 265 seconds]
alp has joined #nixos-dev
Synthetica has joined #nixos-dev
ris has joined #nixos-dev
<Synthetica> What's going on with r-ryantm? There haven't been any pulls in almost 2 weeks: https://github.com/nixos/nixpkgs/pulls?q=is%3Apr+author%3Ar-ryantm
ris has quit [Ping timeout: 246 seconds]
orivej has joined #nixos-dev
niksnut has joined #nixos-dev
avn has quit [Ping timeout: 246 seconds]
avn has joined #nixos-dev
<MichaelRaskin> I remember a change towards doing full nixpkgs-review on each PR, maybe this lead to some more infrastructure work than initially done?
__monty__ has joined #nixos-dev
<adisbladis> It seems like it got stuck on a stalled build
<adisbladis> I've restarted the service
FRidh has joined #nixos-dev
<Synthetica> adisbladis: Awesome :)
<hyperfekt> since most packaging happens by example i think we could do extremely well in providing an index of packaging styles and packages that implement them 🤔
<hyperfekt> for example i found no example for an update script for a cargo-built rust package but i'm sure it exists somewhere in nixpkgs
<michaelpj> hyperfekt: as with all such things, the question becomes "who will write this?" and more importantly "who will maintain it?". Will people who create new styles keep this guide up to date?
alp_ has joined #nixos-dev
alp has quit [Read error: Connection reset by peer]
alp_ is now known as alp
<hyperfekt> let's see if we can get something going. i'm writing out a few interesting packages but there's no way i can get through all of them
alp has quit [Remote host closed the connection]
<arianvp> im trying to build the linux kernel; but at the final step of the build it just gets stuck for hours
<arianvp> https://gist.github.com/arianvp/658c58f86a8e7d31581e401d3cdbe812 it's just chilling at 0 CPU for over an hour now at this step
<arianvp> whilst usually I could build the kernel in like 20 minutes I think
alp has joined #nixos-dev
<garbas> Is anybody working on this? Currently the Amazon EC2 tab looks empty -> https://github.com/NixOS/nixos-org-configurations/issues/109
<{^_^}> nixos-org-configurations#109 (by knl, 6 days ago, open): update AWS AMIs and Virtual Box Images to 20.03
teto has joined #nixos-dev
<garbas> currently the infra team was doing it but as per https://discourse.nixos.org/t/what-is-the-process-for-publishing-the-official-nixos-amis/4581/2?u=knl i think proposal is that release team will do it
<gchristensen> the infra team has to do the images I think
<gchristensen> but the images were already made and the list was applied to 20.03, so not sure why it isn't updating on the website yet?
<niksnut> maybe flake.lock needs to be updated?
<gchristensen> quite probable
<hyperfekt> This could serve as an intermediary between 'add instructions to the manual' and 'reinvent the wheel for every package': https://nixos.wiki/wiki/Packaging_Examples
<garbas> niksnut: gchristensen: looking at the problem. looks like update.sh in nixos-homepage is not working (or at least failing for us to notice)
* garbas first needs to make sure update.sh fails before trying to fix it
phreedom has quit [Ping timeout: 240 seconds]
phreedom_ has joined #nixos-dev
phreedom_ has quit [Quit: No Ping reply in 180 seconds.]
phreedom has joined #nixos-dev
<flokli> euurgh, staging currently got broken by the binutils/coreutils bump: https://github.com/NixOS/nixpkgs/pull/85951#issuecomment-621199816 (what arianvp was asking earlier about)
<gchristensen> owwww
<flokli> and I still don't know why `ac_cv_func_lchmod=no` fixed it halfway, but now rsync still is broken…
<flokli> low level stuff…
<flokli> https://bugzilla.redhat.com/show_bug.cgi?id=1811038#c4 looks suspicious - maybe the nix sandbox is blocking some of this?
<gchristensen> wow
<LnL> well that fixed it for coreutils but the change was binutils
<flokli> well, but if I revert the binutils bumps, rsync doesn't fail anymore
<flokli> so something from inside binutils must change how rsync behaves here.
<gchristensen> probably should
<LnL> had a feeling "fixing" coreutils wasn't a good idea, I bet rsync isn't going to be the only other thing that suffers from whatever changed
<gchristensen> +1
<flokli> LnL: any idea what part of binutils might involve coreutils' and rsync's behaviour here?
<flokli> if we can fix it in there, we don't need to patch half userland.
<flokli> (or investigate how to make it work in the sanbox, at least not immediately)
<LnL> I have no clue, just reverting binutils and waiting for upstream to fix it seems like a better idea
<flokli> according to fedora (https://bugzilla.redhat.com/show_bug.cgi?id=1811038#c12), "upstream" might be glibc here
<flokli> but this doesn't make any sense
<flokli> yes, of course, we can revert, but I'm not aware if any "upstream" so far is aware of it
<flokli> so, building rsync with the same configureFlags hack gets rsync to work, and the kernel to build
<flokli> I'll post something in the fedora bugtracker and hope to hear something back
<flokli> and yeah, if I don't find anything quickly, best thing might be to revert for now
<LnL> yeah, I'd point them at our issue and the fact that binutils might be involved
<LnL> good chance we're noticing this earlier because we've built everything using this new binutils
FRidh has quit [Ping timeout: 256 seconds]
<infinisil> ,random-pr
<{^_^}> https://github.com/NixOS/nixpkgs/pull/76427 (by countoren, 18 weeks ago, open): vscode-utils/vscodeEnv: add vscodeWithConfiguration, vscodeExts2nix a…
alp has quit [Ping timeout: 272 seconds]
<{^_^}> #86324 (by flokli, 1 minute ago, open): staging: revert the binutils bump
CRTified has quit [Quit: Gateway shutdown]
ixxie has joined #nixos-dev
cole-h has joined #nixos-dev
alp has joined #nixos-dev
justanotheruser has quit [Ping timeout: 265 seconds]
johnny101 has quit [Read error: Connection reset by peer]
johnny101 has joined #nixos-dev
justanotheruser has joined #nixos-dev
drakonis has joined #nixos-dev
FRidh has joined #nixos-dev
evanjs has quit [Quit: ZNC 1.7.5 - https://znc.in]
evanjs has joined #nixos-dev
ixxie has quit [Quit: leaving]
ixxie has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
<elvishjerricco> Ugh, I'm getting very annoyed by the nuke-refs in initrd. I feel like I'd rather not use makeInitrd, considering we're just nuking refs manually anyway. It'd be easier to just copy the files into the archive in their normal positions and avoid having to patch them
ris has joined #nixos-dev
<gchristensen> meh. nix doesn't even support attribute names containing a thorn. how am I supposed to declaratively manage my midieval Scandinavian machines.
<gchristensen> oh I can quote it. neat
<cole-h> (Sorry, but: medieval)
<cole-h> (Unless your machines produce music)
<gchristensen> good catch :)
<danderson> now I want nix to produce MIDI output during evaluation
<danderson> "man, this machine sounds wrong, I must have missed a piece of config"
<cole-h> Midi-nix? Midnix? Minix? Wait a second...
<gchristensen> iirc there was a log-message-to-midi thing so you could "hear" your production system
<clever> gchristensen: i once had cron configured to beep a different tone every hour, 15mins, and 30mins
<MichaelRaskin> I think there was something for code, too
<clever> gchristensen: the same pc was also configured to beep at 15khz any time my name was said on irc, well above my dads hearing, so he couldnt complain when it beeped non-stop for an hour :P
<MichaelRaskin> Musical profiling of evaluation performance is closer to how debugging on MCUs happen
<gchristensen> hah!
<danderson> I've debugged an LCD driver using audio before
<danderson> turns out it's easy to output register values with dits and dahs
<clever> danderson: ive seen a blog about dumping the ipod mask rom via the piezo beeper
<clever> danderson: because they could run custom code, but didnt know how to drive the lcd or get serial
<danderson> heh, yeah, it's a cool technique. Needs a computer receiving stuff if you want to send lots of data though
<danderson> I was just dumping a couple bytes at a time
nschoe has joined #nixos-dev
<arianvp> is the nix.distributedBuilds option in nixos working?
<arianvp> trying to set up remote builds but cant get it to work
<arianvp> looking at the nix-daemon.nix code it seems like it's missing a line "builders = @/etc/nix/machines"
<arianvp> /etc/nix/machines is generated but never passed to nix as an option :/
<clever> arianvp: what does `nix show-config | grep builders` say?
<arianvp> ah it does show @/etc/nix/machines
<clever> arianvp: that would be the default value for when it isnt in nix.conf
<arianvp> but but when I run nix-build --max-jobs 0 nothing happens :/
<arianvp> it just says "please enable remote building"
<arianvp> and I dont see nix-daemon even trying to ssh into the box (no traffic)
<clever> arianvp: what is the contents of the machines file?
<arianvp> arian@ryzen.local x86_64-linux /root/.ssh/id_ed25519 8 1
<clever> arianvp: and what arch are you building for?
<arianvp> x86_64-linux
<clever> arianvp: try building with -vvv?
<arianvp> got 1 remote builders
<arianvp> considering building on remote machine 'ssh://@/etc/nix/machines'
<arianvp> hook reply is 'decline'
<clever> arianvp: which version of nix?
nschoe has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<arianvp> 2.3.4
<arianvp> whatever 20.03 ship
<arianvp> seems the include syntax got broken :/
<clever> ive got 2.3.1 here and it works just fine
<arianvp> are you on 20.03?
<clever> nixos-unstable, but havent updated recently
phreedom has quit [Remote host closed the connection]
<arianvp> i'll have another look later. thanks for trying to help
phreedom has joined #nixos-dev
<arianvp> worked aorund by scheduling a build like this:
<arianvp> nix build -f . linuxPackages.kernel --store ssh-ng://arian@ryzen.loca
<arianvp> seems a bit more elegant
<clever> arianvp: that wont copy the product back to your local machine
<arianvp> oh :(
<clever> it does the entire build on the remote box
<clever> you want --option builders 'something'
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
<arianvp> I tried nix-build -A linuxPackages.kernel --builders ssh://arian@ryzen.local --max-jobs 0
<arianvp> got 1 remote builders
<arianvp> considering building on remote machine 'ssh://arian@ryzen.local'
FRidh has quit [Ping timeout: 260 seconds]
<arianvp> hook reply is 'decline'
<arianvp> same issue again it seems
<arianvp> (and nix-daemon isn't even ssh'ing into the box; I see no traffic)
FRidh has joined #nixos-dev
prusnak has joined #nixos-dev
<clever> arianvp: try with an ip, not .local?
ixxie has quit [Ping timeout: 244 seconds]
<arianvp> will try when home again
ixxie has joined #nixos-dev
<arianvp> clever: also didn't work
<arianvp> same result. just says "hook reply is 'decline'"
<adisbladis> Re yesterdays subuid/subgid/containers discussion I'd like some feedback on https://github.com/NixOS/nixpkgs/pull/86278
<{^_^}> #86278 (by adisbladis, 16 hours ago, open): nixos.users-groups: Set up subuid/subgid mappings for all normal users
<infinisil> Ah you're doing it at runtime now
<adisbladis> Yeah.. I admitted defeat
<adisbladis> (Defeat looks a lot like perl)
<cole-h> Why are some things in perl, others in bash/sh, and yet others in python? "Just because"?
<adisbladis> cole-h: Reasons ¯\_(ツ)_/¯
<MichaelRaskin> Because some people think that Perl is more readable than Bash.
<adisbladis> I think we should at least be semi-consistent about what goes in the minimal installer
<MichaelRaskin> And for some time Python was not in the mandatory closure
<cole-h> Alright, that makes sense.
<adisbladis> Python still has a pretty significant disadvantage over Perl for these type of scripts (and I'm saying this as someone who does a lot of python)
<adisbladis> Python startup performance is _terrible_
<qyliss> Is Perl still bad at cross-compiling?
<adisbladis> qyliss: iirc yes
<adisbladis> An advantage with Python is that we can get the closure much smaller than Perl, with more in stdlib so we don't need to pull in modules for JSON parsing and such
<adisbladis> perl == 84.2M
<adisbladis> python3Minimal == 54.1M
infinisil has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
infinisil has joined #nixos-dev
<adisbladis> It'd be interesting to reimplement all of that in a compiled language
<cole-h> Am I hearing "RiiR"? :P
<qyliss> what if we rewrote it all in Scheme :P
<adisbladis> <3 qyliss
<{^_^}> qyliss's karma got increased to 55
teto has quit [Quit: WeeChat 2.8]
ixxie has quit [Quit: Lost terminal]
FRidh has quit [Quit: Konversation terminated!]
<MichaelRaskin> Well, if we want simple compilation, reasonable compiled size and readability, that would sound like (modern) Pascal
<MichaelRaskin> But I agree that rewriting it in Scheme has that benefit of not needing to actually write
<emily> just rewrite it all in execline. people don't need to be able to read these things right?
infinisil has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
infinisil has joined #nixos-dev
infinisil has quit [Client Quit]
infinisil has joined #nixos-dev
<worldofpeace> so I've made an acme team for nixpkgs https://github.com/orgs/NixOS/teams/acme/members, but for some reason I'm a member as "maintainer"
drakonis has quit [Quit: WeeChat 2.8]
<cole-h> You're the "maintainer" of the team because you created, I guess?
<worldofpeace> I guess that makes sense, I probably should read github docs on teams
<cole-h> The real question is if their docs are up to date, or accurate.
<arianvp> still no luck in getting remote builders working on 20.03. Anybody else experiencing issues with it?
* cole-h glares at GH sending status updates without the single required field, `state` , to ofborg fairly often
teto has joined #nixos-dev
<ma27[m]> arianvp: both my remote machine and my local machine are on 20.03 and I don't have any issues. Not sure I can help, but what's happening?
<arianvp> no matter what I do I always get:
<arianvp> got 1 remote builders
<arianvp> considering building on remote machine 'ssh://192.168.0.23'
<arianvp> hook reply is 'decline'
<arianvp> and nix-daemon isn't even trying to connect to the builder
<arianvp> it declines before even trying; not srue what it means
<arianvp> same happens when using builder = @/etc/nix/machines instead
<LnL> have you setup known_hosts for root?
<LnL> alltho I error messages should be better for that now
<arianvp> root has already ssh'd into the box; yes
<gchristensen> try: ssh 192.168.0.23 nix-store --version
<arianvp> # ssh arian@ryzen.local nix-store --version
<arianvp> nix-store (Nix) 2.3.4
<arianvp> (Again; nix-daemon doesn't even get to the point where it executes the ssh command as far as I can see in htop or in my sshd logs)
<arianvp> maybe rightType = false in this branch? but not sure what that would mean
<arianvp> Oh I think I figured it out...
<arianvp> linuxPackages.kernel.requiredSystemFeatures
<arianvp> [ "big-parallel" ]
<arianvp> :/
<arianvp> I think the remote build is silently failing because the remote builder doesn't have this feature
<clever> arianvp: nix should be giving an error, saying what features it wants
<arianvp> oh
<arianvp> then it's not that. still no clue then
<clever> arianvp: you can still try adding it to the /etc/nix/machines file though
<ma27[m]> ahh.... can't you force a remote build using `-j 0`?
<arianvp> ma27[m]: yes with --max-jobs 0 ' that's what I'm doing
<clever> ma27[m]: the only machine in /etc/nix/machines lacks the big-parallel flag, so its not a viable one, and then it fails claiming you need to add remote builders
<gchristensen> is arian the user who nix will be building as?
<gchristensen> sshing as*
<arianvp> Ok adding the features got me a step further
<arianvp> (we should really have an error message for this though? is this a bug?)
<clever> arianvp: the user you ssh into, must be trusted on the remote machines nix.conf
<arianvp> ah
<arianvp> that's an easy fix
<ma27[m]> are you building with a user who's allowed to build on the server?
<arianvp> but nix not showing an erro when the features mismatch sounds like a bug right? I think it used to do that before
<arianvp> it's remote building now! \o/
<clever> arianvp: yeah, that would be a bug, one min...
<clever> src/libstore/build.cc: throw Error("a '%s' with features {%s} is required to build '%s', but I am a '%s' with features {%s}",
<arianvp> thanks for the help folks
<clever> arianvp: it will say this, when the features dont match up right
<clever> arianvp: id say the error you got was worse and very misleading
<arianvp> i'll open a bug report tomorrow
<arianvp> im kind of sleep-deprived so maybe I read over the error 10 times. let me recheck tomorrow :P
justanotheruser has quit [Ping timeout: 264 seconds]
drakonis has joined #nixos-dev
__monty__ has quit [Quit: leaving]
justanotheruser has joined #nixos-dev
alp has quit [Ping timeout: 265 seconds]
Synthetica has quit [Quit: Connection closed for inactivity]