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 | 18.03 release managers: fpletz and vcunat
kmicklas has quit [Ping timeout: 255 seconds]
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm has joined #nixos-dev
Lisanna has joined #nixos-dev
Lisanna has quit [Remote host closed the connection]
pxc has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 245 seconds]
FRidh has joined #nixos-dev
MichaelRaskin has quit [Quit: MichaelRaskin]
<thoughtpolice> Ugh. postgis is a postgresql extension, which depends on gdal, which can optionally depend on postgresql itself in order to do queries against postgis inside postgresql. o_O The net effect of this is that postgresql93 depends on postgresql96, lol
<thoughtpolice> Or well, if you build postgis against postgres93, you will build postgres96, to fulfil the default 'postgres' dependency gdal has
<thoughtpolice> what a chore
zybell_ has quit [Read error: Connection reset by peer]
zybell_ has joined #nixos-dev
orivej has joined #nixos-dev
__Sander__ has joined #nixos-dev
ma27 has joined #nixos-dev
ma27 has quit [Client Quit]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 255 seconds]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 268 seconds]
NinjaTrappeur has quit [Changing host]
NinjaTrappeur has joined #nixos-dev
ma27 has joined #nixos-dev
<Mic92> it's time for s/python3/python/ otherwise redhat will overtake us.
<Mic92> Also its more psychological/political thing.
<niksnut> Mic92: we decided a while ago not to do that
<Mic92> ah ok
goibhniu has joined #nixos-dev
__Sander__ has quit [Ping timeout: 260 seconds]
zybell_ has quit [Ping timeout: 256 seconds]
<LnL> oh? not do that or not do that for now
<srhb> Is there anything we can do to bump the 18.03 tests higher up? The zfs data loss bugfix still hasn't gone into the channel.
<LnL> that doesn't sounds scary
zybell_ has joined #nixos-dev
<srhb> I don't know how likely it is that anyone hits it, but ick.
<LnL> bumped the jobs to the front of the queue, that might help
<srhb> LnL: Thank you. :)
__Sander__ has joined #nixos-dev
ma27 has quit [Ping timeout: 260 seconds]
ma27 has joined #nixos-dev
vdemeester` has quit [Changing host]
vdemeester` has joined #nixos-dev
jtojnar has joined #nixos-dev
<srhb> LnL: 72665983 of tested went through, thank you very much. :)
MoreTea2 is now known as MoreTea
phreedom has quit [Ping timeout: 268 seconds]
<layus> There is currently a bit of churn on updating git to 2.17 because of the changes in how git builds perl
<layus> Just look at git log --graph --oneline --all pkgs/applications/version-management/git-and-tools/git/default.nix
<layus> The update has already been commited three times then reverted.
phreedom has joined #nixos-dev
<LnL> oh, yours wasn't the first pr?
<layus> apparently some peaple tried before on staging, and were reverted too.
<LnL> :/
<layus> We all missed the perl update.
<layus> Now that https://github.com/NixOS/nixpkgs/pull/38763 has the perl update, we should test and merge it to staging before we see even more attempts.
<LnL> oh ok so just the same issue
<LnL> I didn't even know that some of the commands depend on perl
<LnL> thought it where the hooks
<layus> apparently, the inbteractive commands (git add -p/-i, and some more) require perl
jtojnar_ has joined #nixos-dev
<layus> Also, master branch seems to fail g*ofborg evaluation ATM.
jtojnar has quit [Ping timeout: 268 seconds]
jtojnar_ is now known as jtojnar
<gchristensen> for 45 minutes, layus
<layus> gchristensen, can I help with that ?
<gchristensen> sure, check out the evaluation failure, and then look to see what commit from the last hour introduced it
<LnL> I'm at work so can't commit
<gchristensen> from there, drop a note on the commit that broke it, and either fix it (ideally) or if you don't know how to fix it, revert it if the original author doesn't fix it within ~10min or so. I'm also not able to fix it right now.
<gchristensen> then comment "@grahamcofborg eval" on every PR that failed due to master being broken
<layus> call me stupid, but I am unable to find the command that ofBorg uses to evaluate most of the packages.
<gchristensen> well so an easy way is to just look at the Details link on a PR which has failed
<gchristensen> click Details on the failed check
<layus> ok, I am bootstrapped now :-)
<gchristensen> awesome, thank you :D
<layus> Yes, it is the culprit?
<layus> *!
<layus> Now the question is why (and potentially revert it after ~10min if I understood correctly) :-)
<pbogdan> I _think_ https://github.com/NixOS/nixpkgs/commit/78d5690ddc23e7a98f32b8554d279156e2e2ecd0#diff-1d34614b675a9e766a1baf2fd4848968R99 is supposed to say platforms.all (rather than just platforms)
<jtojnar> Michael Raskin claims to have fixed it but I do not see any commit
<samueldr> would it be wise to add, in the contributions guide, the command to run to evaluate nixpkgs so that a dev, locally, hopefully sees the failure?
<samueldr> (just like how nox review is in the guides)
<gchristensen> something like that might be reasonable
<gchristensen> but the whole thing takes quite a bit of RAM to run (8G), so may not be reasonable for them to do the whole thing
<samueldr> evaluation is safe, right? though is it long?
<samueldr> ah right
<LnL> we could add nix-instantiate '<nixpkgs>' --eval --strict -A hello.meta as a check to the list
<layus> The fix of Michael Raskin works indeed. Still seeing no commit. Should I submit one ?
<layus> I created https://github.com/NixOS/nixpkgs/pull/38776. Could someone merge it ?
<layus> I guess that the evaluation failure is because master is evaluated too right ? because it works for me locally
<gchristensen> yeah, you'll just have to merge it
<gchristensen> or I can
<layus> I can't. I am not powerfull enough ;-)
<gchristensen> https://github.com/NixOS/nixpkgs/pull/38775 let's see if this passes
<shlevy> :o staging merge
<shlevy> We did it... we finally did it
__Sander__ has quit [Ping timeout: 260 seconds]
<layus> gchristensen, if monitoring.nix.ci is the real stuff, it may be worth changing the repo description on https://github.com/NixOS/ofborg
<gchristensen> great idea
FRidh has quit [Remote host closed the connection]
<gchristensen> updated layus
jtojnar_ has joined #nixos-dev
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos-dev
jtojnar_ has quit [Ping timeout: 264 seconds]
ekleog has quit [Quit: WeeChat 2.0]
ekleog has joined #nixos-dev
<shlevy> gchristensen: very slight confusion on the status page... (0 active) next to "broken evaluator" made me think for a second that there were 0 active evals
<shlevy> Not sure there's a good way to avoid that, just FYI
<gchristensen> best thing I can do is delete that page
<gchristensen> unfortunately
goibhniu has quit [Ping timeout: 276 seconds]
<gchristensen> "OfBorg Evals" shows the backlog
<shlevy> Thanks
<Mic92> oh, where have setup.cfg been all the time?
<Mic92> is it possible to generate setup.cfg automatically?
kmicklas has joined #nixos-dev
kgz has quit [Ping timeout: 264 seconds]
<shlevy> Mic92: setupcfg2nix does it for you :)
<shlevy> I need to update the README
<shlevy> but try --help :)
<shlevy> gchristensen: What does it mean when all of ofborg's statuses are green *except* the top level one?
<shlevy> Is there more coming or something stuck?
<gchristensen> its still working
<shlevy> Ah OK
<gchristensen> maybe if there are >1 items, it should always have 2 of them or 0 of them pending
<LnL> what's the general opinion on this? https://github.com/NixOS/nixpkgs/pull/38791
<domenkozar> LnL: we did the same for python long time ago
<domenkozar> I think it's hard to say if it's good or bad
<domenkozar> you can see things should be built with purity (nix-shell --pure or sandboxing) so it doesn't matter to set these
<LnL> did you see the comment I linked
<LnL> it's not just building things
<domenkozar> I see, it affects runtime
<domenkozar> yeah, that's why we removed it for python, there's more inconvienience than convenience
<domenkozar> and it's better to have it opt-in
<domenkozar> if we're doing this purity thing :)
<LnL> and that also means that nix run might not be using the perl libraries you expect
<LnL> even if it doesn't fail
<LnL> yeah, the pipefail was tricky because all of the output was hidden
<LnL> wrong channel
Sonarpulse has joined #nixos-dev
<domenkozar> shlevy: oh nice, PEP426!!
<shlevy> domenkozar: :)
<domenkozar> when I slowly backed off from python community it was still WIP after ~5 years
<domenkozar> is it accepted yet / implemented fully?
<shlevy> Not sure, there's enough of it for Nix :)
<domenkozar> well the problem is if packages don't use it, then it's not much use :)
<shlevy> I've seen a few packages here and there
<shlevy> But we can insist on it for teams we support ;)
<shlevy> And setupcfg2nix can try to generate one from a setup.py https://github.com/target/setupcfg2nix/blob/master/setupcfg2nix/cli.py#L15-L38
<domenkozar> uhh
<domenkozar> don't get me started why that won't work :)
<shlevy> Certainly won't work everywhere
<shlevy> But it's worked for a handful of random test cases ;)
<shlevy> The goal is to make manual nix wrangling less likely, not eliminate it :)
<domenkozar> I think python should move to similar state as haskell
<domenkozar> settiling on spec
<shlevy> Haskell hasn't settled on a spec either ;)
<domenkozar> then writing setup.py -> setup.cfg conversion
<shlevy> cabal's spec is "source code" :D
<domenkozar> and nix should only consume the latter
<shlevy> But yes, agreed
<domenkozar> shlevy: well cabal is non-turing complete
<domenkozar> compared to setup.py
<shlevy> The --allow-autogeneration flag is only there for use in development workflows, definitely not usable for nixpkgs itself
<domenkozar> so it is speced, although not documented
<shlevy> Sure :)
<domenkozar> it's like nixops, do you really want to maintain this layer of constant pain?
<domenkozar> many have tried, wasted years, realized it should be fixed upstream, got lost in python packaging politics and rage quit
<domenkozar> maybe I'm wrong and pep426 is legit now
<domenkozar> let me check
<gchristensen> pretty sure it was domenkozar who talked me out of packaging and maintaining 1password-in-wine for the same reason -- constant pain.
<domenkozar> shlevy: if it's just a "getting things done" thing that's fine
<domenkozar> just wanting to give you a heads up what you're up against
<domenkozar> hope it's helpful :)
<shlevy> domenkozar: Well at the very list all of the Target packages supported by Nix will use setup.cfg :)
<domenkozar> so this is in final draft
<domenkozar> which replaces PEP426
<domenkozar> 6 years and counting :)
<shlevy> :)
<domenkozar> shlevy: sounds good :)
<domenkozar> shlevy: great work btw getting this effort started
<domenkozar> it has even longer history :)
<shlevy> :D
<thoughtpolice> gchristensen: Mmmm, do you know what ever happened to #36386?
<thoughtpolice> gchristensen: As part of my postgresql work I did these updates myself and found your PR later! I'd be happy to commit them if you want; things seem clean to me though I can do a once over. Not sure what was blocking it originally...
<gchristensen> I think I just forgot about it
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-dev
zybell_ has quit [Ping timeout: 265 seconds]
zybell_ has joined #nixos-dev
<shlevy> dtz: I'm getting an error ending with "NOTE: A missing vtable usually means the first non-inline virtual member function has no definition." when trying to build the new nix-plugins on darwin. Is this the vtable anchoring thing?
<zybell_> y
pie__ has joined #nixos-dev
pie_ has quit [Read error: Connection reset by peer]
pie_ has joined #nixos-dev
pie__ has quit [Read error: Connection reset by peer]
<dtz> haha yes
<dtz> :(
angerman_ has joined #nixos-dev
thoughtpolice_ has joined #nixos-dev
MichaelRaskin has joined #nixos-dev
florianjacob has quit [*.net *.split]
adisbladis[m] has quit [*.net *.split]
angerman has quit [*.net *.split]
thoughtpolice has quit [*.net *.split]
thoughtpolice_ is now known as thoughtpolice
angerman_ is now known as angerman
adisbladis[m] has joined #nixos-dev
florianjacob has joined #nixos-dev
pie__ has joined #nixos-dev
pie_ has quit [Read error: Connection reset by peer]
<shlevy> dtz: What's the right way to fix it? I'm building plugins so they're not linked (or they shouldn't be)...
guest____ has joined #nixos-dev
<zybell_> You should look for the first virt func that you *do* implement, and order it first of virtuals in header/decl. If you don't own the header,you can lookup the first func and implement a dummy for it.
kragniz has joined #nixos-dev
ma27 has quit [Ping timeout: 245 seconds]
infinisil has quit [Quit: Configuring ZNC, sorry for the join/quits!]
infinisil has joined #nixos-dev
infinisil has quit [Client Quit]
infinisil has joined #nixos-dev
infinisil has quit [Client Quit]
guest____ has left #nixos-dev [#nixos-dev]
infinisil has joined #nixos-dev
jtojnar_ has joined #nixos-dev
jtojnar has quit [Read error: Connection reset by peer]
jtojnar_ has quit [Remote host closed the connection]
jtojnar has joined #nixos-dev
<ekleog> suggestion: revert the dovecot 2.3.0.1 -> 2.3.1 update on release-18.03
<ekleog> I'm getting Panic: file message-address.c: line 147 (parse_angle_addr): assertion failed: (*ctx->parser.data == '<') since the change between commits b6ddb9913f2..489a14add9a
<ekleog> (with dovecot's deliver-agent)
<gchristensen> who updated it?
<ekleog> ryantm's bot
<LnL> but that's the release branch
<gchristensen> who backported it?
<ekleog> and I can't see who cherry-picked it : 2ddacd295940910bf561cdafb04f38b69352e7f3
<ekleog> even git show shows only ryantm
<ekleog> fpletz already fixed another issue in 69a03de0384d767a155a459c5385e2b316578bc0, though
<ekleog> oh, dotlambda backported, github's webui appears better than git show here https://github.com/NixOS/nixpkgs/commit/2ddacd295940910bf561cdafb04f38b69352e7f3
<Dezgeg> --pretty=fuller is needed to display that
<ekleog> (oh didn't know, always assumed git show defaulted to show everything, thanks!)
<gchristensen> should probably talk to dotlambda and see why it was backported
<ekleog> hmm anyone knows if dotlambda is ping-able on IRC? I can't find the rationale, https://www.dovecot.org/doc/NEWS shows no security issue -- unless it was assumed unlikely to break things, I guess my opensmtpd+rspamd+dovecot setup isn't really the most tested one by dovecot upstream
<gchristensen> dotlambda is in #nixos
<ekleog> ok thanks! :)
<gchristensen> in general though a 2.3.0 -> 2.3.1 patch isn't something I'd hesitate too long on backporting, so not sure it was strictly wrongg
<ekleog> yeah I don't think it was wrongdoing, just wondering whether it could/should be reverted :)
<gchristensen> yeah
<gchristensen> maybe simultaneously go bark up the dovecot tree and see if they know about it
<ekleog> yeah, just did that :°
<gchristensen> awesome :)
ekleog has quit [Quit: back soon]
ekleog has joined #nixos-dev
{^_^} has quit [Ping timeout: 260 seconds]
kragniz is now known as kgz
<Sonarpulse> need +1
<gchristensen> done :)
<Sonarpulse> gchristensen: thanks!
<dtz> \o/
<shlevy> By the way, turned out the vtable thing was cmake not passing the right linker flags for plugins
{^_^} has joined #nixos-dev
{^_^} has quit [Changing host]
{^_^} has joined #nixos-dev
<shlevy> And now, a teaser of things to come https://twitter.com/shlevy/status/984205196212473858
<shlevy> copumpkin: ^ :)
<copumpkin> ooh
<shlevy> extraBuiltins via https://github.com/shlevy/nix-plugins, project containing fetch-pypi-hash now going through the Target open sourcing process
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos-dev
orivej has quit [Ping timeout: 246 seconds]
pie__ has quit [Read error: Connection reset by peer]
pie_ has joined #nixos-dev
<jtojnar> should we merge https://github.com/NixOS/nixpkgs/pull/38696 to staging as well?
<shlevy> jtojnar: We can merge master forward to staging at any time
<jtojnar> cool
<jtojnar> should I do a pull request or can I just merge it locally and then push
<shlevy> Just merge master in and push
<jtojnar> ok, done
jtojnar_ has joined #nixos-dev
jtojnar has quit [Ping timeout: 240 seconds]
jtojnar_ is now known as jtojnar