gchristensen changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | 18.09 release managers: vcunat and samueldr | https://logs.nix.samueldr.com/nixos-dev
<samueldr> neat
<samueldr> hey, about that -aarch64 jobset...
<gchristensen> oh
<samueldr> :)
<samueldr> tyvm
<gchristensen> ?1
<gchristensen> +1
sir_guy_carleton has joined #nixos-dev
<samueldr> hey all, ('morning for those reading after their sleep cycle)
<samueldr> how useful would something like that be for ZHF?
lassulus_ has joined #nixos-dev
lassulus has quit [Ping timeout: 240 seconds]
lassulus_ is now known as lassulus
drakonis has joined #nixos-dev
drakonis has quit [Remote host closed the connection]
sir_guy_carleton has quit [Quit: WeeChat 2.0]
orivej has joined #nixos-dev
vcunat has joined #nixos-dev
<vcunat> samueldr: instead of the lists I think it's better to provide a link to hydra.nixos.org where people can see the fresh status
<LnL> depends how many toplevel vs propagated issues there are
<vcunat> why?
Mic92_ has quit [Ping timeout: 252 seconds]
nwspk has quit [Ping timeout: 260 seconds]
<LnL> you can't filter that on hydra
<vcunat> the gist doesn't seem to do that either
<vcunat> (it looks close to a snapshot of the page to me)
Mic92_ has joined #nixos-dev
nwspk has joined #nixos-dev
<LnL> ah
<LnL> ideally hydra could just do it, but I've found the ZHF checkmarks pretty useful even tho they get out of sync
<LnL> I also wonder how hard it would be to make something that can detect simple errors like a missing header file
<vcunat> samueldr: I started staging-18.09, so we can merge such stuff in batches
<vcunat> we may be missing a large fraction of security fixes until people get used to having an 18.09 branch
<vcunat> LnL: I'd love if Hydra could group the dep-failed by a dependency that failed them (not precise, there may be multiple known, but better than now)
<vcunat> now I just "randomly sample" several of them, look and the source of the failure, try to fix it and then wait for fixed re-evaluation, as I'm lazy to click so many links
__Sander__ has joined #nixos-dev
<LnL> oh, that would be even better
<LnL> I was talking about just the red vs green separation
<LnL> gray*
<vcunat> Yes, I know. Sometimes the gray ones don't have corresponding red ones, but usually they do.
<vcunat> Perhaps the separation would be a simple change, even JS-only, but I don't know this stuff.
vcunat has quit [Ping timeout: 252 seconds]
goibhniu has joined #nixos-dev
goibhniu has quit [Ping timeout: 246 seconds]
goibhniu has joined #nixos-dev
__Sander__ has quit [Ping timeout: 245 seconds]
__Sander__ has joined #nixos-dev
<domenkozar> only 421 build failures?
<domenkozar> :O
<domenkozar> gchristensen++ I think we can attribute that a lot to ofborg
* LnL doesn't want to look at the darwin side
<domenkozar> :P
<domenkozar> it's not bad
<domenkozar> 830
<domenkozar> and I'd bet half of that is haskell/python
<LnL> don't remember what it was previously
<domenkozar> ~3k when we started with darwin channel
<domenkozar> 577 on 18.03
<LnL> but maybe I'll get it below 500 this time then
<LnL> yeah, the first release was _really_ rough
vcunat has joined #nixos-dev
<vcunat> domenkozar: your idea of starting ZHF has proved to be a very good one
<vcunat> aarch64 is still over a thousand, but it's the youngest of these platforms
nbp_ has joined #nixos-dev
<aminechikhaoui> samueldr: the gist seems better imho for showing which packages are failing, as hydra is so slow to load the jobs and navigate through
<Dezgeg> yup. also it's useful to collect timestamps when the build started failing to see if it was broken ages ago or never succeeded
<vcunat> https://hydra.nixos.org/eval/1476786 seems to load quite fast for me
<vcunat> (still failing vs. newly failing might be slightly confusing, I suppose)
<aminechikhaoui> also try still failing tab then click (x more builds omitted) at the bottom, that takes too much to load for me
<vcunat> yes, that takes a long time due to loading 50k builds
<vcunat> (all of them, not just failed; count depends on jobset)
<domenkozar> vcunat: someone else started ZHF, I just aligned it with releases :)
<domenkozar> hmm or was it me?
<domenkozar> no
<domenkozar> it was Luca Bruno
<domenkozar> I miss that guy :)
<domenkozar> good times, when I still had lots of time :D
<vcunat> yes, probably him - the apparently oldest ZHF commit: https://github.com/NixOS/nixpkgs/commit/59cfbf4d7
<domenkozar> first ZHF and removal of builderDefsPackage
<domenkozar> that's like a double-karma
<vcunat> :-)
<vcunat> ++lethalman
<yorick> aminechikhaoui: domenkozar: is nixops dead?
<aminechikhaoui> :O
<aminechikhaoui> why ? :)
<yorick> aws deployment has been broken for half a year, none of the PRs are getting merged
<yorick> I suppose 18.09 will fix it
<aminechikhaoui> we use AWS at work all the time, what's broken ?
<yorick> the AMI runs out of memory during setup and never comes up
<aminechikhaoui> is this on a small instance type ?
<aminechikhaoui> is there an issue already for that btw ?
<yorick> yeah. there is a PR for it, even
<yorick> but I think everyone started maintaining their own forks and wrappers
<aminechikhaoui> the out of memory is a Nix issue, I believe the new Nix release should improve things, did you try with Nix 2.1 ?
<yorick> aminechikhaoui: there is no AMI with nix 2.1 yet
<yorick> so the current workaround is deploying the 17.09 AMI and then upgrading it to -unstable
<vcunat> :-)
<domenkozar> yorick: no, but nixops was always maintained with some delays :)
<domenkozar> it should sort out with Nix 2.1
<aminechikhaoui> I guess we can create some AMIs with 2.1
<yorick> I got permission to spend 30% of my work time on nixos stuff. maybe you want more maintainers? ;)
<aminechikhaoui> not sure if it's propagated already to the channels
<vcunat> no, I don't think so
<vcunat> It was pushed today to master
<aminechikhaoui> oh so not even in 18.03 yet ?
<yorick> 2.1 got released today
<aminechikhaoui> cool
<yorick> oh man, that would fix so much
<aminechikhaoui> fingers crossed :)
<yorick> currently it's "here's the production network, it only comes up in singapore, meh"
<aminechikhaoui> btw is there an expression I can test with later ? or does it happen with just setting t2.micro for example ?
<yorick> (I suppose singapore network latency is high enough that it doesn't run out of memory)
<yorick> aminechikhaoui: yeah. also, t3.micro is out now
<yorick> t3.nano, even
<aminechikhaoui> alright thanks
<vcunat> you can just pick that commit e.g. as an overlay - there should be almost no rebuilds
<yorick> ec2-properties.nix is outdated but the genertaor script is broken
<domenkozar> yorick: any reviews and contributions are welcome :)
<infinisil> domenkozar: Hey, could you have a look at https://github.com/domenkozar/hie-nix/pull/26 ?
<{^_^}> domenkozar/hie-nix#26 (by Infinisil, 1 week ago, open): Update nixpkgs, stack2nix and HIE
<domenkozar> yeah will do :)
<domenkozar> could switch to circleci
<infinisil> Thanks :D
<domenkozar> but it's still an issue since PRs like that don't have access to signing key
<domenkozar> maybe circleci has a solution :)
* infinisil has no idea about CI's
<LnL> vault provides a reasonable way to handle this kind of stuff IMHO
<yorick> https://github.com/NixOS/nix/issues/1756 the bane of my existence and a single line fix
<{^_^}> nix#1756 (by dtzWill, 36 weeks ago, open): "nix repl" cannot tab-complete names that contain a dash ('-')
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos-dev
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos-dev
<srhb> Is there any way to link to a section of the NixOS manual from an option description? I think not (there's some kind of stage problem, I think) -- but checking anyway :)
<gchristensen> srhb: hmmm ... I think so ... one sec
<manveru> might be good to fix all of those links sometime :) https://repology.org/repository/nix_unstable/problems
<vcunat> Well, pushed the fix, seeing no use to wait with this.
<manveru> thanks :)
drakonis has joined #nixos-dev
<nbp_> gchristensen: Indeed, I do not mind about the section ids.
nbp_ is now known as nbp
<gchristensen> nbp_: thank you :)
<samueldr> I read the comments, I'll keep doing the gists, and try to improve it with regards to propagated failures
__Sander__ has quit [Quit: Konversation terminated!]
vcunat has quit [Ping timeout: 250 seconds]
drakonis has quit [Remote host closed the connection]
vcunat has joined #nixos-dev
vcunat has quit [Quit: Leaving.]
vcunat has joined #nixos-dev
xeji has joined #nixos-dev
<xeji> gchristensen: ofborg builds still don't seem to work, no auto builds: #45976, no manual triggering: #45966 #45967 #45968 (even with only one space :). Builds may be started but results don't come back to github.
<{^_^}> https://github.com/NixOS/nixpkgs/pull/45976 (by rnhmjoj, 5 hours ago, open): imv: 2.1.3 -> 3.0.0
<{^_^}> https://github.com/NixOS/nixpkgs/pull/45966 (by romildo, 15 hours ago, open): qt5dxcb-plugin: init at 1.1.11
<{^_^}> https://github.com/NixOS/nixpkgs/pull/45967 (by worldofpeace, 13 hours ago, open): Various build fixes
<gchristensen> hmmmmm
<gchristensen> hmmm
<gchristensen> ...hmmm...
<gchristensen> oh dear!
<xeji> oops
<xeji> at least they didn't get lost :)
<gchristensen> it posted so many comments it got temporarily blocked, and maybe 4-5 got lost
<gchristensen> still 20 more to post :/
<LnL> oh
<gchristensen> gonna wait 5min and start it again
<xeji> great, thanks! Doesn't github have a process to increase API limits for bots?
<gchristensen> yeah
<gchristensen> this isn't an API Limit but an abuse limit
<xeji> resistance is futile...
<gchristensen> "You have triggered an abuse detection mechanism and have been temporarily blocked from content creation. Please retry your request again later."
<LnL> stop starting so many builds!
<LnL> :p
<gchristensen> for the best, what if ofborg was in a hard loop? :)
<LnL> hmm yeah, can rabbitmq funnel a queue or would this need to be something in the poster?
<gchristensen> funnel?
<gchristensen> like rate-limit?
<LnL> yeah limit what can go through based on time/amount
<gchristensen> oh not sure
<LnL> and buffer the rest
<gchristensen> ok we're all caught up now, thank you for the ping again xeji
<xeji> thanks for your help!
<xeji> gchristensen: sorry to trouble you again. I can't view any recent ofborg logs, like #45976 and others. ofborg-viewer gets stuck at 'Subscribing to "nixos/nixpkgs.45976"'.
<{^_^}> https://github.com/NixOS/nixpkgs/pull/45976 (by rnhmjoj, 6 hours ago, open): imv: 2.1.3 -> 3.0.0
<gchristensen> heh
<gchristensen> fixed
<xeji> hmmm, doesn't work for me yet
<gchristensen> it'll only work for logs going forward
<xeji> ok, thx
yorick has quit [*.net *.split]
bgamari has quit [*.net *.split]
ocharles has quit [*.net *.split]
pauldub has quit [*.net *.split]
taktoa[c] has quit [*.net *.split]
ghuntley has quit [*.net *.split]
zimbatm has quit [*.net *.split]
manveru has quit [*.net *.split]
thoughtpolice has quit [*.net *.split]
niksnut has quit [*.net *.split]
phreedom_ has quit [*.net *.split]
niksnut has joined #nixos-dev
phreedom has joined #nixos-dev
sir_guy_carleton has joined #nixos-dev
jtojnar has quit [Remote host closed the connection]
jtojnar has joined #nixos-dev
drakonis_ has joined #nixos-dev
drakonis_ has quit [Remote host closed the connection]
vcunat has quit [Quit: Leaving.]
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos-dev
orivej has quit [Ping timeout: 244 seconds]
goibhniu has quit [Ping timeout: 246 seconds]
xeji has quit [Quit: WeeChat 2.1]