worldofpeace changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | NixOS 20.09 Nightingale ✨ https://discourse.nixos.org/t/nixos-20-09-release/9668 | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | https://r13y.com | 20.09 RMs: worldofpeace, jonringer | https://logs.nix.samueldr.com/nixos-dev
<jtojnar> thank you
ehmry has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
ehmry has joined #nixos-dev
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-dev
stigo has quit [Ping timeout: 246 seconds]
ashkitten has quit [Quit: WeeChat 3.0]
stigo has joined #nixos-dev
ashkitten has joined #nixos-dev
orivej has quit [Ping timeout: 256 seconds]
supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos-dev
ris has quit [Ping timeout: 246 seconds]
rajivr has joined #nixos-dev
LnL has quit [Quit: exit 1]
LnL has joined #nixos-dev
LnL has joined #nixos-dev
LnL has quit [Changing host]
justanotheruser has joined #nixos-dev
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 264 seconds]
evils has joined #nixos-dev
pmy has quit [Ping timeout: 264 seconds]
pmy has joined #nixos-dev
lopsided98 has quit [Ping timeout: 264 seconds]
lopsided98 has joined #nixos-dev
lopsided98 has quit [Ping timeout: 264 seconds]
jonringer has quit [Ping timeout: 264 seconds]
lopsided98 has joined #nixos-dev
cole-h has quit [Ping timeout: 265 seconds]
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-dev
lopsided98 has quit [Ping timeout: 264 seconds]
__monty__ has joined #nixos-dev
lopsided98 has joined #nixos-dev
orivej has joined #nixos-dev
Mic92 has quit [Quit: WeeChat 3.1]
Mic92 has joined #nixos-dev
<das_j> what was the URL of the nixpkgs PR viewer?
<das_j> I think it was by alyssa
<siraben> Is it just me or nixpkgs-review is breaking on latest master?
<siraben> 546 packages removed:
<siraben> vscode-extension-redhat-vscode-yaml (†0.13.0) vscode-extension-rubbersheep-gi (†0.2.11) vscode-extension-Rubymaniac-vscode-paste-and-indent (†0.0.8) vscode-extension-rust-analyzer-unstable (†2021-03-15) vscode-extension-ryu1kn-partial-diff (†1.4.1) ...
<sterni> maybe an eval error which is tripping nix-env up?
<siraben> ttribute 'version' missing
<siraben> attribute*
<siraben> can anyone else confirm?
<sterni> seems like #116941 is the culprit then
<{^_^}> https://github.com/NixOS/nixpkgs/pull/116941 (by rhoriguchi, 2 days ago, merged): vscode-extensions.redhat.java: init at 0.76.0
<sterni> siraben: are you on macOS? maybe jdk isn't available there
<siraben> Yeah I'm on macOS.
<siraben> jdk is there but doesn't have a version attribute
<das_j> huh? Where is the merge event for that PR?
<sterni> siraben: pkgs/development/compilers/openjdk/darwin/default.nix
<sterni> siraben: jdk only has name = "zulu15.28.51-ca-jdk15.0.1";
<siraben> Yeah, sigh.
<siraben> (this is why we'll need to totally remove use of name)
<__monty__> Ironic how badly named name is : )
<sterni> siraben: nope, not everything is versioned
<siraben> Or as much as possible
<sterni> also user derivations don't need to be versioned
<sterni> and stuff like writeText
<siraben> #117214
<{^_^}> https://github.com/NixOS/nixpkgs/pull/117214 (by siraben, 6 seconds ago, open): openjdk/darwin: move version out of name
<sterni> but yeah for packages it'd be nice
<siraben> also would be good to have more darwin maintainers, as always
<sterni> or drop darwin :p
<sterni> whyyyyyyyy do NixOS test script need to please some auto formatter
<siraben> noooo don't drop darwin it's way better the homebrew
<siraben> lol
<siraben> s/the/than
<sterni> these are autogenerated scripts and I just wanna use string interpolation in nix expressions without jumping through a thousand hoops
<siraben> it just needs more .app applications packaged and such
<sterni> we should've stuck with a non whitespace sensitive language like perl for those in hindsight :p
<siraben> lol nixpkgs-review says "250 packages updated:"
<siraben> hopefully it doesn't build them
<siraben> must be all the java packages we have
<sterni> you bot it's gonna
<supersandro2000> siraben: yeah, noticed the same with your PR before that
<siraben> well shit fetching is taking forever, uh, ^C
pmy has quit [Quit: WeeChat 3.1]
<siraben> in zulu15.28.51-ca-jdk15.0.1, which part is the version string?
<supersandro2000> sterni: because you should not write whatever you want
<supersandro2000> siraben: 🤷
<supersandro2000> 15.28.51-ca-jdk15.0.1?
<siraben> jdk.version seems to work now btw
<siraben> I let the version string be 15.0.1
<supersandro2000> yeah did the same fix in another PR a few minutes before you
<siraben> Oh I see you closed it
<sterni> supersandro2000: the script is _automatically_ generated
<sterni> but I have to fucking check that the line doesn't get too long this is ridiculous
<sterni> the resulting script is ever only seen be the python interpreter
<symphorien[m]> I thought there was a option to set to escape the formatting check but I can't find it in the manual
<supersandro2000> add a ignore rule to the editor check
<supersandro2000> for normal nix files this makes total sense
<supersandro2000> or let black shorten it
<sterni> supersandro2000: it's not about the editor, but the nixos test literally failing when black has anything to say against it
<supersandro2000> 🤔 we maybe want to soften that
<supersandro2000> blacks line length ideas are questionable
<supersandro2000> maybe we want to get rid of that
<supersandro2000> black really likes to add multiple line breaks to [ or ( just to shave of two characters
<supersandro2000> short term idea: run black on the file after generation
<sterni> black is executed twice???
<sterni> supersandro2000: it is done after generation, but who cares about that nobody reads that file
<sterni> and black can't format nix source anyways
<supersandro2000> generate file -> black $file -> run tests?
<supersandro2000> yeah, I first thought you are talking about editor check but black there makes not to much sense IMHO
<sterni> oh you can pass skipLint but it's a warning, so probably trips up ofborg
Raito_Bezarius has quit [Ping timeout: 264 seconds]
<sterni> #117219
<{^_^}> https://github.com/NixOS/nixpkgs/pull/117219 (by sternenseemann, 8 seconds ago, open): nixos/lib/testing-python: default to skipLint = true
pmy has joined #nixos-dev
pmy has quit [Client Quit]
<siraben> Could I be added to NixOS/darwin-maintainers?
<gchristensen> yeah, I don't think that makes sense sterni, use black to reformat your generated script if that is the issue
<sterni> gchristensen: that would require ifd?!
<sterni> although maybe not
<sterni> testScript can't be a derivation so I'd have to at least call builtins.readFile on a derivation
<sterni> which hydra wouldn't like right?
<gchristensen> gotcha, yeah, that is tricky -- what about your test requires generation?
<__monty__> Wouldn't it just be a postBuildHook?
<gchristensen> maybe you can write your test to a file then import it as a module and call functions
<sterni> gchristensen: I have multiple machines which are supposed to execute the same stuff in the same sequence
mkaito has joined #nixos-dev
mkaito has quit [Changing host]
mkaito has joined #nixos-dev
<sterni> so I have a lib.concatStrings and a bit of stuff
<sterni> and I do string interpolation so I don't copy everything everywhere
<gchristensen> could you use python's loop constructs to implement it by hand, without generation?
<sterni> and I use nix variables instead of python ones so I can reuse them in the machine config
<sterni> yeah sure, but I can't share the stuff with the machine configuration
<gchristensen> have an attrset that gets written out as json and loaded in the test?
<sterni> yeah I guess that could work
<sterni> but it's a bit annoying to have to jump through hoops for that when the idiomatic way would be to just do string interpolation
<sterni> at least I feel like it is :p
<gchristensen> yeah, understood
<gchristensen> it isn't great, but I'm really really not excited to loosen format requirements
<sterni> hm okay but this I can't do I think:
<supersandro2000> siraben: side question: Did you already take a look at the stdenv?
<sterni> I have a list of the client machine names in nix I also use for generating their configurations
<sterni> I need to reuse them in python
<sterni> but there they are variable names so I can't really pass them around or pass by value?!
<sterni> okay but probably works with some mild-ish string interpolation
<sterni> but now I need to convert a nix list into a well formatted python list with variables inside it still
<sterni> okay different issue: what do I need to do to have user services in tests?
<sterni> # Failed to connect to bus: No such file or directory
<siraben> supersandro2000: the uutils stdenv?
<supersandro2000> siraben: the darwin one
aranea is now known as mewra
jonringer has joined #nixos-dev
<siraben> supersandro2000: could you link to it?
<siraben> i've only looked through linux stdenv
<siraben> maybe i'll look into darwin in a bit
<siraben> specifically I looked into how to create a very minimal stdenv smaller than stdenvNoCC
pmy has joined #nixos-dev
pmy has quit [Client Quit]
pmy has joined #nixos-dev
pmy has quit [Client Quit]
pmy has joined #nixos-dev
Raito_Bezarius has joined #nixos-dev
pmy has quit [Quit: WeeChat 3.1]
pmy has joined #nixos-dev
<gchristensen> :)
<sterni> apparantley it doesn't like an empty line at the end of the file
<sterni> wait maybe I can fix it without that hack
pmy has quit [Quit: WeeChat 3.1]
Baughn has quit [Ping timeout: 246 seconds]
orivej has quit [Ping timeout: 244 seconds]
pmy has joined #nixos-dev
orivej has joined #nixos-dev
das_j has quit [Ping timeout: 240 seconds]
ajs124 has quit [Ping timeout: 240 seconds]
Scriptkiddi has quit [Ping timeout: 265 seconds]
das_j has joined #nixos-dev
Scriptkiddi has joined #nixos-dev
ajs124 has joined #nixos-dev
lukegb is now known as lukeg
Baughn has joined #nixos-dev
<s1341_> is there a way to access a funciton in a let section of another derivation? I want to reuse the makeLinuxHeaders funciton here: https://github.com/NixOS/nixpkgs/blob/master/pkgs/os-specific/linux/kernel-headers/default.nix
jonringer has quit [Remote host closed the connection]
jonringer has joined #nixos-dev
<__monty__> s1341_: No let really makes a local scope. Though I wouldn't put it past #nix-lang to find a way to circumvent that.
<sterni> s1341_: nope, but it is in the set anyways, maybe you could just expose it in all-packages.nix as well?
<s1341_> sterni: how would i do that...
<s1341_> ?
<sterni> s1341_: in all-packages.nix there is a inherit (callPackage ../os-specific/linux… {}) linuxHeaders; you could just add makeLinuxHeaders there and PR it if you think that is a good idea
<s1341_> ok. i will do so.
<s1341_> thanks.
<s1341_> sterni: while i have your attention, i'm trying to track down why xorgserver depends on systemd.
<s1341_> why-depends shows me that Xephyr et al. have systemd in their rpath
<s1341_> because they link with dbus....
<s1341_> but I disabled dbus in the xorgserver's derivation.
<s1341_> but it still pulls in systemd....
<s1341_> any advice?
<clever> that reminds me, X appears to be broken on my nas, possible all of nixos-unstable, but i dont see how it got thru testing?
<clever> Mar 21 15:43:47 nas sddm[2496]: Logind interface found
<clever> this happens any time i try to start X
Baughn has quit [Ping timeout: 256 seconds]
<sterni> s1341_: at least that what it seems to me like, you'll have to check I have no clue about xorg compilation in nixpkgs in all honestly
<s1341_> i removed that when building for android....
<sterni> *honesty
<s1341_> sterni: how can I modify the ARCH in that makeLinuxHeaders function?
<andi-> likely udev being a dependency which is just an alias of systemd
<s1341_> andi! that was it!
<sterni> s1341_: it statically uses stdenvNoCC, so it's not easily changeable
<sterni> probably would need some refactoring
<sterni> making it either overrideable or taking it as an explicit input
<s1341_> nope andi- that wasnt it...
<s1341_> andi-, sterni: how can I chase down this dependency??
<sterni> maybe check nix-store -q --requisites and look for anything suspicious
<sterni> you can also try playing around with nix-tree
cole-h has joined #nixos-dev
<s1341_> nix-tree shows a direct dependency on nix-shell.
<s1341_> s/nix-shell/systemd/
<s1341_> and the derivation itself really does depend on systemd (if i look at the .drv)
Raito_Bezarius has quit [Ping timeout: 264 seconds]
justan0theruser has joined #nixos-dev
justanotheruser has quit [Ping timeout: 265 seconds]
Baughn has joined #nixos-dev
kcalvinalvin has quit [Quit: ZNC 1.7.4 - https://znc.in]
kcalvinalvin has joined #nixos-dev
ris has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
cole-h has quit [Client Quit]
cole-h has joined #nixos-dev
fhdhhw has joined #nixos-dev
rajivr has quit [Quit: Connection closed for inactivity]
Baughn has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
evils has quit [Ping timeout: 245 seconds]
fhdhhw has quit [Quit: Connection closed]
Baughn has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
cole-h has quit [Client Quit]
cole-h has joined #nixos-dev
cole-h has quit [Client Quit]
cole-h has joined #nixos-dev
Baughn has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-dev
Baughn has joined #nixos-dev
mkaito has quit [Quit: WeeChat 3.1]
<s1341_> how do you guys work on the cross toolchains and binutils... every change requires rebuilding a ton of packages until you get hello build?
<s1341_> is there a trick I'm missing?
<infinisil> s1341_: Are you changing nixpkgs or only the package sources?
<s1341_> infinisil nixpkgs
<infinisil> Hm not sure
<s1341_> :*(
<samueldr> patience *and* fast machines I guess
<s1341_> samueldr: well i have a 200 core machine, and i've set it up as a remote builder, but it's not particularly effective... i only get a handful of builds concurrently and i get stuck with a bottleneck every once in a while...
<s1341_> samueldr: do you have recomendations?
<s1341_> because my cores are idle and want to be fed!
<samueldr> sorry, none, I guess you're better equpiped with fast machines than I was
<samueldr> maybe I had more patience?
<s1341_> perhaps ;)
tomberek has joined #nixos-dev
<sterni> which one is the staging jobset on hydra or is there none?
<sterni> always get lost in that instance
<das_j> I bet 4 days for nixpkgs to get it to master
<das_j> sorry that should have gone to nixos-chat…
<gchristensen> in preparation for that we should resolve the issues keeping the releases back: https://status.nixos.org/
<gchristensen> can anyone take point on that?
<das_j> ajs told me there is already a fix for the unstable-small blocker
<das_j> the jdk stuff
<gchristensen> in master?
<das_j> no as a PR
<das_j> umm where is it
<sterni> yeah I'm on it
<sterni> it's by siraben, but I'm rebasing it currently, I think I can get it to 0 rebuilds
<sterni> which would be good for unblocking the channel
<{^_^}> #117214 (by siraben, 9 hours ago, open): openjdk/darwin: move version out of name
<sterni> I'm a bit confused as to why this is happening in unstable-small though maybe there's another issue
<sterni> since unstable-small doesn't care about darwin?!
<sterni> ah yeah it does test x86_64-darwin
<sterni> I'm gonna wait for ofborg and if it's no rebuilds squash and merge
<sterni> we can change the name attribute some other time I guess, but it triggers > 250 rebuilds due to changed outpaths so might not be the best idea rn
<supersandro2000> 250 rebuilds is nothing
<supersandro2000> this will be done in no time
<supersandro2000> it is an eval issue that only occurs when you eval the vscode-extensions.redhat.java
<supersandro2000> because darwin override jdk to some weird version which is still using name
<supersandro2000> and rebuild reduction here is just a waste of time.
<supersandro2000> also because version gets added it still causes a rebuild. reverting that workaround and squash merging that
<sterni> ok doesn't work without rebuilds
<cole-h> sterni: just fyi, but your attempt to remove rebuilds won't work because `pname` and `version` are part of the derivation
<supersandro2000> cole-h: thanks. tried to say that earlier
<sterni> cole-h: yeah I remembered then :)
<sterni> cole-h: input hash changes unfortunately
<supersandro2000> also how should someone know that a channel is blocked if they don't check hydra daily?
<cole-h> gchristensen: There should be no more channel-blocking issues now. But the failure in https://hydra.nixos.org/build/139585985 should already have been fixed a while ago...
Raito_Bezarius has joined #nixos-dev
<sterni> supersandro2000: https://status.nixos.org/ is a bit neater
<supersandro2000> we can't expect authors of PRs to check https://hydra.nixos.org/job/nixos/unstable-small/tested if a PR they did broke some channel
<supersandro2000> sterni: well the graphs show me that both channel are broken for a while
<supersandro2000> if I happen to find the blue link on the bright green bars I might also know the time period
<sterni> you can click on the build problem thing and get to see what the issue is relatively easily
<sterni> trunk-combined and unstable-small both failed because of the darwin thing it seems
<cole-h> oh I guess the unshare issue hasn't been fixed it: https://github.com/NixOS/nixpkgs/pull/114602
<{^_^}> #114602 (by kvtb, 3 weeks ago, open): nix: fix "common.sh: line 94: unshare: command not found"
<supersandro2000> I am inclined to throw that page into huginn and send me a message when the channel is blocked/broken
<sterni> a hydra irc bot would be neat but it probably is not really feasible
* cole-h goes to force-push that PR since the author is unresponsive...
<sterni> either too noisy or not useful
<supersandro2000> I would have fixed that eval issue on darwin yesterday if I would have known about it
<supersandro2000> but I didn't because I did not check status.nixos.org in the last months
<supersandro2000> sterni: why not dump it into an extra channel only for that?
<supersandro2000> is 114602 also a channel blocker?
<sterni> supersandro2000: trunk-combined fails
<cole-h> Like, the issue only seems to appear every so often
<sterni> because nixpkgs-review depends on nixUnstable and is a constituent
<sterni> the rebase bot just doesn't like me does it
<supersandro2000> thats the line it is called https://github.com/NixOS/nix/blob/master/tests/common.sh.in#L92
<supersandro2000> rebase action is very flaky lately. No idea why.
<sterni> lol it fails the same way for me every time
<sterni> no error message
<supersandro2000> nix file in nix has a util-linux minimal https://github.com/NixOS/nix/blob/master/flake.nix#L93
<cole-h> yep, testing that right now before pushing
<cole-h> (it also has it in buildInputs rather than nativeBuildInputs)
<sterni> supersandro2000: can you try rebasing #116770 onto staging
<{^_^}> https://github.com/NixOS/nixpkgs/pull/116770 (by r-ryantm, 4 days ago, closed): ocamlPackages.findlib: 1.8.1 -> 1.9.1
<sterni> otherwise I can probably just cherry pick it
<sterni> you can't change the target branch for other ppls prs apparently
<ryantm> sterni: r-ryantm's PRs are committed against a merge base of staging and master, so you should be able to retarget them by clicking edit at the top of the PR and changing the target branch.
<ryantm> It doesn't look like you can do that while it is closed though.
<sterni> ryantm: ohhh
<sterni> ryantm: thanks that does the trick :)
mkaito has joined #nixos-dev
mkaito has joined #nixos-dev
mkaito has quit [Changing host]
<ryantm> sterni: woo. If you do that for most people's PRs it will spam the hell out of everyone, but targeting the merge-base fixes that.
<ryantm> I mean committing on top of the merge base.
<cole-h> what's the nix option that rebuilds packages a certain number of times to test reprodulity?
<cole-h> reproducibility*
<ryantm> cole-h: https://r13y.com/
<cole-h> I mean the command line option
<cole-h> or nix.conf option
<sterni> ryantm: that's pretty cool
<ryantm> scroll down on that page cole-h
<ryantm> --check --keep-failed
<sterni> ryantm: although I've noticed that it's kind of alright if you change target first and force push afterwards
<cole-h> --check only does it once
<cole-h> I thought there was an option that would rebuild N number of times
<ryantm> --option build-repeat N
<ryantm> at least that's something in the manual
<cole-h> ah, --repeat, thanks.
<cole-h> (build-repeat is a deprecated alias)
<ryantm> Ah yeah.
<cole-h> ryantm++
<{^_^}> ryantm's karma got increased to 36
<ryantm> sterni: Whenever I change the target of someone's PR, it notifies like 30 people and shows hundreds of commits on the PR. Well, this is based on information from years ago because I would never try it again.
<sterni> ryantm: wait I got it wrong
<sterni> ryantm: you need to rebase onto staging first and then change the target branch that way you avoid having the thousands of new master commits in your PR temporarily
<sterni> but it also has a good chance of pinging a few people by accident I guess
<samueldr> you can do it locally using `git merge-base` to find the common ancestor, and rebase on that
<samueldr> then changing the pr target branch will not ping
<ryantm> Right, that's what r-ryantm does.
<samueldr> changing the target OR pushing, in either order, has the same probabilities of pinging the whole world
<samueldr> as long as the merge base is not the same
<ryantm> So the only hope is to target the merge-base from the start?
<gchristensen> or force push
<samueldr> exactly, rebase on merge-base, force push, then you're free to change as you want
<sterni> you guys sound like professionals
<ryantm> sterni: Grats on being a comitter now.
<sterni> ryantm: thx :)
<sterni> this comes at the perfect time … *checks notes* when I really should start writing that paper
<cole-h> jk, I can't force-push that PR, they unticked the box :(
<cole-h> guess I'll open a new PR :)
copumpkin has quit [Remote host closed the connection]
copumpkin has joined #nixos-dev