jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined joined #nixos-dev
MoreTea has quit [(Ping timeout: 240 seconds)]
MoreTea2 has joined joined #nixos-dev
bgamari has quit [(Ping timeout: 240 seconds)]
pbogdan has quit [(Ping timeout: 240 seconds)]
pbogdan has joined joined #nixos-dev
catern has joined joined #nixos-dev
bgamari has joined joined #nixos-dev
phreedom has joined joined #nixos-dev
<gchristensen> I'm afraid multi-user nix on mac being the default is not long for the world
<gchristensen> creating users takes more than just being root
<gchristensen> you either have to call: sysadminctl interactive -addUser <- and this prompts for the admin password every time it is called
<gchristensen> or ... you call sysadminctl -adminUser <username> -adminPassword <password> -addUser <- but here you have to prompt the username/password ahead of time and pass it in over the CLI, which feels icky
<gchristensen> for multi-user support perhaps a better approach is focusing on an upgrade path
<gchristensen> that way, users who want more safer nix on darwin can choose to, and suffer through the password prompts for a few minutes and not be scared away when they're first using it
<gchristensen> I feel bad that this is the case and I would love there to be an alternative, but I don't seee one
<gchristensen> maybe someone else knows some wizardry
mbrgm has quit [(Ping timeout: 248 seconds)]
mbrgm has joined joined #nixos-dev
jtojnar has quit [(Remote host closed the connection)]
jtojnar has joined joined #nixos-dev
orivej has quit [(Ping timeout: 240 seconds)]
page_ has joined joined #nixos-dev
page_ has quit [(Changing host)]
page_ has joined joined #nixos-dev
page has quit [(Ping timeout: 268 seconds)]
vcunat has joined joined #nixos-dev
adisbladis has quit [(Changing host)]
adisbladis has joined joined #nixos-dev
vcunat has quit [(Quit: Leaving.)]
page_ is now known as page
vcunat has joined joined #nixos-dev
_ts_ has joined joined #nixos-dev
FRidh has joined joined #nixos-dev
<vcunat> niksnut: Hydra's hanging jobs seem to have been blocking nixpkgs-17.09-darwin channel for two weeks now. The jobs are in queue but never get built. Cancelling and/or restarting won't help, as they just re-appear in the queue, waiting for something (a some zombie job I suspect).
goibhniu1 is now known as goibhniu
orivej has joined joined #nixos-dev
avn has quit [(Remote host closed the connection)]
FRidh has quit [(Ping timeout: 248 seconds)]
orivej_ has joined joined #nixos-dev
orivej has quit [(Ping timeout: 258 seconds)]
<_ts_> hi, I wanted to move NixUP forward [https://github.com/NixOS/nixpkgs/pull/9250]. Would it be possible to merge it into staging or something already and to verify none of the tests break?
FRidh has joined joined #nixos-dev
<vcunat> _ts_: it's not a mass rebuild
<vcunat> not at all
<vcunat> I'll create a separate jobset
FRidh has quit [(Read error: Connection reset by peer)]
<_ts_> vcunat, that would be nice, thank you so much!
<vcunat> _ts_: BTW, this weekend I'm likely to finally get to reviewing the C stuff
<vcunat> (as promised)
<_ts_> vcunat, that's great! Thank you very much for investing your time! :)
<vcunat> :-) I don't expect it will need much time, but we'll see.
<_ts_> vcunat: I _hope_ so, but you'll see! ;)
<niksnut> vcunat: what job are you referring to?
<niksnut> jobs
<vcunat> niksnut: if you restart the aborted jobs in https://hydra.nixos.org/eval/1410885
<vcunat> those have been hanging
<vcunat> but I don't know what in particular is blocking them
<niksnut> thanks
<vcunat> niksnut: I suspected some of the jobs that have been "running for days" in https://hydra.nixos.org/machines
<vcunat> I'm not sure what that means precisely
<vcunat> still, there are currently no such jobs shown for darwin in /machines
<niksnut> vcunat: I restarted one of them (https://hydra.nixos.org/build/63291951) and it seems to be building
<niksnut> so I've restarted the others as well
<vcunat> OK, we'll see.
jtojnar has quit [(Remote host closed the connection)]
<vcunat> _ts_: the test failures shown there have been resolved on master, IIRC
<vcunat> so a rebase/merge should fix them
Jackneillll has joined joined #nixos-dev
Jackneilll has quit [(Ping timeout: 240 seconds)]
<vcunat> :-) darwin-tested is a job for x86_64-linux
<vcunat> not that it matters, really; perhaps it's just more difficult to run it locally
<_ts_> vcunat: thanks, I just rebased my PR. Will hydra re-evaluate the jobset automatically?
<vcunat> I forced it now, as the period is rather long
orivej_ has quit [(Ping timeout: 240 seconds)]
<MoreTea2> hm, is pkgs.my_pkg.meta.maintainers actually used anywhere? (e.g. hydra)
MoreTea2 is now known as MoreTea
<clever> MoreTea: i think hydra will use it to email the maintainers when the build fails
<MoreTea> I was under the impression that it was, but now I'm browsing through Hydra's source, and the email notifications appear to be based if you touched (or not) a file in the repository.
<clever> MoreTea: i only get notifications for things where i'm a maintainer, nothing for things ive PR'd but dont maintain
<MoreTea> I'll do some more digging then.
<clever> MoreTea: this is one of the emails i sometimes receive
<clever> MoreTea: line 67 checks the maintainers list of the build
<clever> MoreTea: i think that one is for scanning the build inputs, seperate from the maintainers
<clever> MoreTea: your function is called on line 77 of the file i linked
<clever> MoreTea: so it emails both sets of users
<clever> maintainers always get pinged, so they can clean up after other users, and it sounds like it also checks git history and pings anybody that modified it between the good eval an the bad eval
<clever> MoreTea: i think something in the eval area of hydra will parse the eval state, and generate the maintainers listed here: https://hydra.nixos.org/build/64192951#tabs-details
<MoreTea> yes, I'm looking for that last part.
<clever> so we are trying to find the ->maintainers var in the perl wrappers over the postgress
<clever> MoreTea: i think thats it
<clever> i also see a lot of undocumented hydra features like maxSilent and timeout!
<MoreTea> great, thanks.
<gchristensen> MoreTea: how about ... :) nix-instantiate --json --strict --eval . -A git.meta.maintainers
<MoreTea> gchristensen: that's a simple list of emails.
<gchristensen> sorry, I missed what you were looking for and assumed you were going for that :)
<MoreTea> I want:
<MoreTea> eelco = { GitHub ="edolstra"; email = "Eelco Dolstra <eelco.dolstra@logicblox.com>"; IRC = "niksnut"; };
<MoreTea>
<gchristensen> oh, you want to expand our maintainer list, I see
<vcunat> well, general advice is to keep all the three identifiers the same
<MoreTea> yes, and the assumption that pkgs.my_package.meta.maintainers is a list of strings is probably hardcoded everywhere.
<vcunat> for GitHub names it's even written at the top of the maintainers file
<gchristensen> vcunat: unfortunately not always practical
<vcunat> MoreTea: the list is in lib/maintainers.nix
<MoreTea> vcunat: I know. See the PR I created to have a backward compatible output for e.g. hydra, and structured information for pinging the correct maintainers on GitHub.
<gchristensen> link to your PR?
<MoreTea> The problem is that you can't even access the key names of the attrset when you go through pkgs.my_package.meta.mainatiners; it wil be a list of strings.
<gchristensen> you'd have to do a messy lookup again to get the key
<gchristensen> MoreTea: if eelco is saying just make it structured, I'd just make it structured :P I think that means he's volunteering to fix hydra
<MoreTea> That will take ages
<gchristensen> oh?
<vcunat> right, I see, you can't get the ID nicely from meta
<MoreTea> I mean, this will probably break everyone's hydra, right? So we should not just simply update hydra, and immediately push the structure to lib/maintainers
<gchristensen> hydra is "unreleased software" so I wouldn't worry _too_ much about breaking it, but your caution is noted
<gchristensen> (and perhaps a good idea)
<vcunat> if only the e-mailing feature breaks, that shouldn't be that bad
<MoreTea> Hence the "that will take ages", because then we need to update Hydra first and offer a window for people to upgrade their installations
<niksnut> I don't think we need to be backward compatible here
<gchristensen> I propose we make a branch where we just do it, create a jobset to see just how it breaks hydra and then see what we do
<niksnut> fixing hydra will take 5 minutes
<MoreTea> cool.
<gchristensen> or maybe niksnut knows how it breaks hydra and knows it won't be so terrible for people who use it
<niksnut> I think it will just ignore it
<MoreTea> clever pointed me to where it's actually used.
<gchristensen> perfect!
<MoreTea> So I'll create two PR's then, one for the update of lib/maintainers.nix, and one for hydra ;)
<MoreTea> thanks!
<niksnut> MoreTea: I can fix hydra, it's probably a lot easier for me
<niksnut> unless you want to do some C++ hacking
<clever> there is also the need to update the main hydra when that change gets merged into master
<gchristensen> I'm sure he won't forget to do that :P
<gchristensen> exciting to have more structured maintainers :)
<gchristensen> nice work, thank you, MoreTea, niksnut
orivej has joined joined #nixos-dev
<MoreTea> niksnut: if you're OK with babysitting me, I'd like to excercise my C++ skills a bit :p
FRidh has joined joined #nixos-dev
seppellll has joined joined #nixos-dev
goibhniu has quit [(Remote host closed the connection)]
goibhniu has joined joined #nixos-dev
<copumpkin> it's sort of annoying that nixUnstable is lowPrio
<makefu> everybody is waiting for nixUnstable,a lot of people have already changed because of all the cool features 1.12 provides (not to mention a somewhat sane command line output)
<copumpkin> because there's not an easy way to install 1.12 with that
<copumpkin> there's no standalone installer and nix-env -iA nixpkgs.nixUnstable does waht I'm saying
<makefu> nix.package = pkgs.nixUnstable; works for most parts
adisblad` has joined joined #nixos-dev
<vcunat> niksnut: nixpkgs-17.09-darwin bumped :)
adisbladis has quit [(Ping timeout: 240 seconds)]
adisblad` is now known as adisbladis
adisbladis has quit [(Changing host)]
adisbladis has joined joined #nixos-dev
seppellll has quit [(Ping timeout: 255 seconds)]
seppellll has joined joined #nixos-dev
seppellll has quit [(Ping timeout: 240 seconds)]
<LnL> o/
<LnL> what was the issue
<vcunat> I don't know (and it doesn't seem Eelco knows either)
FRidh has quit [(Quit: Konversation terminated!)]
goibhniu has quit [(Ping timeout: 268 seconds)]
vcunat has quit [(Ping timeout: 240 seconds)]
vcunat has joined joined #nixos-dev
<_ts_> vcunat, I just had a look at the NixUP jobset on hydra, and I'm wondering why there are so may failing tests? Should I have re-based NixUP onto the 17.09 release branch instead of master?
<vcunat> not 17.09
<vcunat> it won't be merged there
<vcunat> > install: missing destination file operand after '/mnt/nix/var/nix/profiles/nixup'
<vcunat> there aren't that many newly failing tests (yet): https://hydra.nixos.org/eval/1410993?compare=1410992#tabs-now-fail
seppellll has joined joined #nixos-dev
<copumpkin> how do people install 1.12 on non-NixOS systems?
<copumpkin> can we just remove the lowPrio on 1.12?
<Dezgeg> whye doesn't just nix-env -iA nixpkgs.nixUnstable work?
<Dezgeg> well you do need to nix-env -e the old nix
<copumpkin> it's awkward because you can't do those in a single atomic operation, so between removing the old nix and installing the new one, you lose nix in your env
<copumpkin> we can work around it but I was just trying to give some colleagues an easy upgrade path
<copumpkin> (like I can save the output of `which nix-env` and then use that after nix-env -e, but ugh)
<Dezgeg> you don't lose it if you install the new one first
<copumpkin> my issue is that someone thought nix-env -iA nixpkgs.nixUnstable worked, except due to the priority it was actually using `nix-*` from 1.11 and `nix` from 1.12
<copumpkin> oh, hmm true
<copumpkin> maybe I'm just being stupid then :)
<copumpkin> let me experiment a bit
tv1 has joined joined #nixos-dev
nixer has joined joined #nixos-dev
tv has quit [(Ping timeout: 268 seconds)]
tv2 has joined joined #nixos-dev
tv1 has quit [(Ping timeout: 268 seconds)]
<_ts_> vcunat, thanks! Could you maybe restart the jobset once more for me, please? I hope I fixed the error.
tv3 has joined joined #nixos-dev
tv2 has quit [(Ping timeout: 240 seconds)]
tv4 has joined joined #nixos-dev
tv3 has quit [(Ping timeout: 240 seconds)]
goibhniu has joined joined #nixos-dev
Sonarpulse has quit [(Ping timeout: 248 seconds)]
seppellll has quit [(Ping timeout: 240 seconds)]
orivej has quit [(Ping timeout: 246 seconds)]
vcunat has quit [(Quit: Leaving.)]
Sonarpulse has joined joined #nixos-dev
<copumpkin> orivej: you're on fire! thanks for handling all those PRs and other things :)
<gchristensen> grahamcofborg will no longer Approve PRs if they build, this is because it may build on linux and fail on darwin, and still leave an Approval. thanks domenkozar for that feedback.
<copumpkin> o/
<Profpatsch> Uh-oh
<Profpatsch> Ah, fixed already? hm
nixer has quit [(Ping timeout: 260 seconds)]
orivej has joined joined #nixos-dev
jtojnar has joined joined #nixos-dev
_ts_ has quit [(Quit: Leaving)]