worldofpeace_ changed the topic of #nixos-dev to: #nixos-dev NixOS Development (#nixos for questions) | NixOS 20.03 BETA Announced | | | 19.09 RMs: disasm, sphalerite; 20.03: worldofpeace, disasm |
abathur has quit [Ping timeout: 256 seconds]
spacekookie_ has joined #nixos-dev
pepesza has quit [Ping timeout: 255 seconds]
ris has joined #nixos-dev
alphawaffle has joined #nixos-dev
betawaffle has quit [Ping timeout: 268 seconds]
spacekookie has quit [Ping timeout: 268 seconds]
multun has quit [Ping timeout: 240 seconds]
Jackneill has quit [Ping timeout: 240 seconds]
LnL has quit [Ping timeout: 240 seconds]
_ris has quit [Ping timeout: 240 seconds]
johnny101 has quit [Ping timeout: 240 seconds]
LnL has joined #nixos-dev
pepesza- has joined #nixos-dev
LnL has joined #nixos-dev
LnL has quit [Changing host]
Jackneill has joined #nixos-dev
multun has joined #nixos-dev
<lovesegfault> gchristensen: ping
<lovesegfault> or Profpatsch
MichaelRaskin has quit [Ping timeout: 258 seconds]
claudiii has quit [Quit: Connection closed for inactivity]
justanotheruser has joined #nixos-dev
lovesegfault has quit [Ping timeout: 256 seconds]
lovesegfault has joined #nixos-dev
justanotheruser has quit [Read error: Connection reset by peer]
<gchristensen> adisbladis: I need this bad enough I'm tempted to make a jobset for it (:
<{^_^}> #82453 (by adisbladis, 1 day ago, open): Python: introduce NIX_PYTHONPREFIX in order to set site.PREFIXES
bhipple has joined #nixos-dev
_ris has joined #nixos-dev
ris has quit [Ping timeout: 272 seconds]
abathur has joined #nixos-dev
lovesegfault has quit [Remote host closed the connection]
lovesegfault has joined #nixos-dev
abathur has quit [Ping timeout: 240 seconds]
justanotheruser has joined #nixos-dev
Emantor has quit [Quit: ZNC -]
Emantor has joined #nixos-dev
orivej has quit [Ping timeout: 256 seconds]
bhipple has quit [Read error: Connection reset by peer]
bhipple has joined #nixos-dev
bhipple has quit [Ping timeout: 268 seconds]
bhipple has joined #nixos-dev
justanotheruser has quit [Ping timeout: 240 seconds]
abathur has joined #nixos-dev
bhipple has quit [Read error: Connection reset by peer]
bhipple has joined #nixos-dev
ash__ has joined #nixos-dev
bhipple_ has joined #nixos-dev
spacekookie has joined #nixos-dev
puckipedia has joined #nixos-dev
clkamp___ has joined #nixos-dev
cptchaos83_ has joined #nixos-dev
arianvp_ has joined #nixos-dev
kraem1 has joined #nixos-dev
Jackneill has quit [*.net *.split]
spacekookie_ has quit [*.net *.split]
andi- has quit [*.net *.split]
cptchaos83 has quit [*.net *.split]
CRTified has quit [*.net *.split]
arianvp has quit [*.net *.split]
puck has quit [*.net *.split]
kraem has quit [*.net *.split]
clkamp_ has quit [*.net *.split]
aminechikhaoui has quit [*.net *.split]
puckipedia is now known as puck
hl has joined #nixos-dev
hl has joined #nixos-dev
hl has quit [Changing host]
aminechikhaoui has joined #nixos-dev
Jackneill has joined #nixos-dev
justanotheruser has joined #nixos-dev
andi- has joined #nixos-dev
ash__ has quit [Quit: ash__]
bhipple_ has quit [Ping timeout: 265 seconds]
bhipple has quit [Ping timeout: 260 seconds]
bhipple has joined #nixos-dev
bhipple_ has joined #nixos-dev
ash__ has joined #nixos-dev
drakonis has quit [Quit: WeeChat 2.7.1]
bhipple_ has quit [Remote host closed the connection]
bhipple has quit [Remote host closed the connection]
abathur has quit [Ping timeout: 268 seconds]
ash__ has quit [Quit: ash__]
<lovesegfault> What's the policy around pkg docs? like with doxygen
<lovesegfault> should they be built?
vcunat has joined #nixos-dev
<vcunat> lovesegfault: I'm not aware of policy. It's often worth separating it into package.doc (if it isn't small).
<lovesegfault> vcunat: got it
<vcunat> Sometimes the bigger issue is excessive blow in build-time closure.
<vcunat> (pandoc or some other tools)
<lovesegfault> ugh, pandoc
<lovesegfault> this one is just doxygen and builds quickly
<vcunat> Yes, probably OK.
<lovesegfault> :+1:
<vcunat> At least for the worst cases the best way I know is making it a separate mkDerivation (if the build system supports building just docs easily).
<vcunat> You can even do that mostly transparently by assigning the derivation to `passthru.doc`
<lovesegfault> I don't really know how to use passthry
<lovesegfault> *passthru
ixxie has joined #nixos-dev
cole-h has quit [Ping timeout: 268 seconds]
<vcunat> That's simple. It's an attrset inside the package definition (stdenv.mkDerivation argument). The attributes inside are available as package.attribute but the build process itself does not have access to them (and thus does not need to be redone).
<vcunat> Well, at least it's very simple to use, that is :-)
<lovesegfault> Hmmm
<lovesegfault> How does that work?
<jtojnar> IIRC, its content is added to the resulting attrset of calling `derivation` function
<jtojnar> so it is not passed to the derivation but it is still part of the result of `mkDerivation`
<vcunat> Yes, that's so. The result is just another attribute set that you can manipulate arbitrarily.
<lovesegfault> Oh, I see
<lovesegfault> that makes sense
<jtojnar> if you want to dig in
* lovesegfault digs in
* lovesegfault learns about extendDerivation
FRidh has joined #nixos-dev
MichaelRaskin has joined #nixos-dev
<vcunat> gchristensen, niksnut: channels are still not updating.
<vcunat> Interesting thing is that now that the conditions for nixos-unstable update started to hold (after a week or so), the updating service for this channel switched to "failed" state.
<danderson> how do I update hackage-packages.nix? I'm removing a package from nixpkgs, and it's part of one of the hackage dependencies
<danderson> but the header of hackage-packages.nix just says it's autogenerated, but not how :(
<vcunat> danderson: I think it's explained here:
<vcunat> Which I discovered now by looking into the manual:
<danderson> ah, I should rtfm more. Thanks.
<vcunat> When you open a PR or issue on GitHub for this, I'd mention at least @peti who seems to be doing most of the regeneration
<danderson> ack, thank you!
<vcunat> :-)
<vcunat> gchristensen, niksnut: would it be difficult to auto-publish all logs from the channel-updating process? It seems not rare that we run into this, and just seeing the error messages might be enough to fix many of these problems. (I don't expect publishing these would be some security risk or something.)
<vcunat> Maybe I feel affected more than most due to doing nix* stuff mainly over weekends.
<MichaelRaskin> Just switch to master
<vcunat> Actually, gating on all those tests and having binaries... is nice.
vcunat has quit [Ping timeout: 255 seconds]
<niksnut> channels are getting updated, see
<niksnut> I think updates are disabled for <reasons>
orivej has joined #nixos-dev
CRTified has joined #nixos-dev
__monty__ has joined #nixos-dev
vcunat has joined #nixos-dev
<vcunat> I thought I did look at earlier today, but perhaps it was only yesterday.
<vcunat> Anyway, thanks.
alphawaffle has quit [Quit: Oh noes, my ZNC!]
betawaffle has joined #nixos-dev
vcunat has quit [Ping timeout: 255 seconds]
<timokau[m]> Is the failure also responsible for the howoldis failure?
ixxie has quit [Ping timeout: 255 seconds]
orivej has quit [Ping timeout: 268 seconds]
<Profpatsch> lovesegfault: If you can build the docs, do it and split them into the "doc" output, yes
<Profpatsch> Also, manpages!
<Profpatsch> With pandoc we can kinda assume that it is already available because many packages share it.
<genesis> hey Profpatsch i'm looking to add checkinput for shellcheck in pkgs/build-support/substitute/
<genesis> should i add some kind of eval "$checkPhase" in it ?
<genesis> changing stuff in this file make a bunch of rebuild btw
<genesis> for the moment i did postInstall = ''
<genesis> '';
<genesis> ${shellcheck}/bin/shellcheck $out/bin/
<genesis> since substitute/substitute-all.nix is a stdenvNoCC.mkDerivation, i wonder why my checkPhase not applied.
orivej has joined #nixos-dev
<{^_^}> (by 3noch, 3 years ago, open): Crazy idea: run ShellCheck on scripts, mkDerivation, runCommand, etc.
<Profpatsch> genesis: I guess you could put it in preCheckPhase?
<Profpatsch> genesis: doCheck is false by default
<genesis> i tried to doCheck=true , let's try preCheckPhase
<Profpatsch> Be aware that shellcheck is a haskell Package, so you might be running in some dependency cycles
<Profpatsch> *into
<makefu> not sure who would be the best to contact, but is currently 404ing
<genesis> Profpatsch : nothing with preCheckPhase
<genesis> neither with installCheckPhase
<Profpatsch> genesis: It uses a different builder than the stdenv builder
<Profpatsch> which isn’t calling any of the usual hooks
<genesis> yes, the pkgs/build-support/substitute/
<genesis> that's why i think we have to add something for checkinputs/checkphase but i let for more nixpkgs maintainers.
<genesis> i keep my postInstall that's not too bad i think.
* emily bumps again; looks like there's consensus to merge at this point?
<{^_^}> #82006 (by emilazy, 6 days ago, open): linuxPackages_{,_latest,_testing}_hardened: enable 32-bit emulation
<emily> thanks gchristensen ^^
<gchristensen> :) thank you
<adisbladis> FRidh: Do you want any help hacking on that python stuff?
<adisbladis> Also building Pypy packages is broken, I have no idea why
<FRidh> adisbladis: if i recall pypy packages are broken because of a different site-packages. At some point I added prebuilt pypy interpreters in order to iterate quicker, but those are broken now as well
<FRidh> Would be good if you could review the python tests. Maybe there's more we can add?
<adisbladis> FRidh: Doing that as we speak :>
<adisbladis> LGTM
<adisbladis> I can't think of anything else right now.
<FRidh> good, thanks
<FRidh> adisbladis: started working on the following to get rid of and perform the steps earlier.
<{^_^}> #82582 (by FRidh, 50 seconds ago, open): Python: patch interpreter to handle NIX_PYTHONHOME
<gchristensen> oh nice, I was talking to someone who was worried site-customize was a bad idea
<FRidh> doubt they would find the proposed solution a better idea :P
<adisbladis> FRidh: Just for vanity reasons I'd like to get in :P
<{^_^}> #82453 (by adisbladis, 1 day ago, open): Python: introduce NIX_PYTHONPREFIX in order to set site.PREFIXES
<FRidh> adisbladis: I suppose you will not be setting NIX_PYTHONPREFIX outside of python.buildEnv?
<gchristensen> FRidh: oh :) I sure wish they'd take a look.
<adisbladis> FRidh: Looks really nice so far!
<adisbladis> FRidh: No, that wouldn't make much sense, would it?
<FRidh> adisbladis: okay, good. Well, I started using NIX_PYTHONPATH :) I think we should go ahead with it, but do update the tests because this should fix one of them
<FRidh> nixenv-venv should also work then, I think
<adisbladis> FRidh: Wdyt about some "real world" tests ?
<FRidh> what do you have in mind?
<adisbladis> That was my minimal repro for the NIX_PYTHONPATH pull request
<FRidh> Right. Yes, I think such tests would definitely be good, it's super hard to keep track of all the things we've broken and fixed
<adisbladis> FRidh: Should I just feel free to tack things onto your PR ?
<adisbladis> Or maybe I'll make a PR against your branch
<{^_^}> #82582 (by FRidh, 15 minutes ago, open): Python: patch interpreter to handle NIX_PYTHONHOME
<adisbladis> It's easy enough to change once we get there
<adisbladis> FRidh: I'll rebase #82453 on staging and update the tests!
<{^_^}> (by adisbladis, 1 day ago, open): Python: introduce NIX_PYTHONPREFIX in order to set site.PREFIXES
ixxie has joined #nixos-dev
abathur has joined #nixos-dev
hexa- has quit [Quit: WeeChat 2.7.1]
hexa- has joined #nixos-dev
abathur has quit [Ping timeout: 240 seconds]
cole-h has joined #nixos-dev
justanotheruser has quit [Ping timeout: 256 seconds]
vcunat has joined #nixos-dev
vcunat has quit [Quit: Leaving.]
abathur has joined #nixos-dev
bhipple has joined #nixos-dev
justanotheruser has joined #nixos-dev
<adisbladis> FRidh: Ready for review ^
<adisbladis> I cherry picked the commit from your comment
<adisbladis> All tests (except pypy which is not yet fininshed) are passing
<genesis> i wonder if we have some build-support for freedesktop menu
<genesis> for the moment in the package i maintained, i need to do that
claudiii has joined #nixos-dev
<genesis> but it would make sens if the menu was inherited for all stuff in a nixpkgs categorie.
drakonis has joined #nixos-dev
disasm has quit [Quit: WeeChat 2.0]
lambdaclock has joined #nixos-dev
disasm has joined #nixos-dev
lambdaclock has quit [Quit: WeeChat 2.7.1]
lambdaclock has joined #nixos-dev
lambdaclock has quit [Client Quit]
lambdaclock has joined #nixos-dev
ixxie has quit [Ping timeout: 256 seconds]
bhipple has quit [Remote host closed the connection]
ixxie has joined #nixos-dev
lambdaclock has quit [Quit: WeeChat 2.7.1]
<ryantm> FRidh: Is it always safe to merge into staging, or is there some kind of merge window for it?
<jtojnar> ryantm It is safe
<jtojnar> the builds are running from staging-next
jtojnar_ has joined #nixos-dev
jtojnar_ has quit [Read error: Connection reset by peer]
jtojnar_ has joined #nixos-dev
jtojnar_ has quit [Read error: Connection reset by peer]
<abathur> gchristensen: ping?
<Profpatsch> abathur: best to just ask your question
<Profpatsch> maybe somebody else can help
<abathur> we need to talk about something I've been picking at that he expressed interest in previously; he asked me to give him a few days and poke him again
<Profpatsch> ah, in that case … ;)
<abathur> I've had no luck poking him via private, but I've seen him respond to pings here
<abathur> so ;)
justanotheruser has quit [Ping timeout: 265 seconds]
justanotheruser has joined #nixos-dev
<garbas> hey! if anybody is interested in helping out i'm trying to organize a team to look primarialy over website and maybe even going further then that ->
<gchristensen> garbas++
<{^_^}> garbas's karma got increased to 3
<samueldr> garbas++
<{^_^}> garbas's karma got increased to 4
<cole-h> garbas++
<{^_^}> garbas's karma got increased to 5
<drakonis> wonderful.
<drakonis> garbas++
<{^_^}> garbas's karma got increased to 6
* garbas hopes that he wont be talking to himself after some time :)
<drakonis> well, you certainly won't
<drakonis> is there a place for the marketing meetings?
<cole-h> Just curious: will these meetings be "open" like the office hours ones? E.g. can a Joe Shmoe like me tune in to hear what's being talked about, without having to actually participate or be knowledgable?
<drakonis> ^
<drakonis> i'd actually like to chime in as there's things worth discussing
<drakonis> there's a distinct lack of demos right now
<samueldr> lack of demos of?
<drakonis> regarding nix functionality
<drakonis> you'd have to take the plunge to see it working
<drakonis> it can be a video or a online application
<drakonis> a script even, just to show it to people so they can see it working rather than simply taking it on face value
<FRidh> I've been thinking of creating jupyter notebooks
<drakonis> it is particularly useful as part of presentations as well
<garbas> cole-h: drakonis: i'm not sure. personally i wouldn't like that this meetings would be "open". but having a writeup afterwords would definetly be provided. i'm also always ready to reconsider.
<garbas> cole-h: drakonis: i would definetly like to limit the discussion to people who are commited and not just people dropping in whenever there is a topic they care about
<cole-h> I meant I wanted to take on more of a role as onlooker than participant. But a post-meeting writeup might be good as well
<samueldr> garbas: nothing stops us from having closed and open meetings, as need be
<garbas> i'll be honest, i'm not sure of a format. at one point i'm "scared" we will bikeshed too much if we open it and on the other hand i'd like to have everybody contributing.
<drakonis> it would also be worthwhile to display how useful it is for specific domains to entice more users
<danderson> one possibility: a livestream format. Have the meeting with specific people, but also livestream it for curious people. That way you're not inviting commentary from random passers-by, but we can still peek in and learn about what's going on.
<genesis> i dreamed about a simple banner with nix logo and nix install instruction for when we do nix packaging
<danderson> (obvious downside, it's more work to set up)
<drakonis> the office hours already have that format
<domenkozar[m]> the fact there's three people from tweag and the role is to promote consulting is quite demotivating to be honest :)
<gchristensen> I would happily step down, then
<samueldr> garbas: you work at tweag?
lovesegfault has quit [Quit: WeeChat 2.7.1]
<garbas> domenkozar[m]: hope you don't think roles are permanent. graham and eelco have enough work on their hands already. but i think having them initially on the team is a good thing.
<garbas> domenkozar[m]: remember i also asked you to join :)
<garbas> samueldr: yeah, just started few weeks ago.
<samueldr> ah, grats on the job then :)
<genesis> to not promote insider-drama, i propose to join
<samueldr> have you read the announcement? details about joining when/if interested are in there :)
<cole-h> danderson: That's what I was trying to imply :P
<drakonis> how long has this been cooking up now?
<drakonis> a week?
<genesis> "connect with companies that offer Nix/NixOS consulting" hum
<genesis> lol
<domenkozar[m]> gchristensen: no need to, but I think we should re-align goals
<domenkozar[m]> my mission is to bring Nix mainstream
<gchristensen> same
<danderson> to provide a counterpoint to "lol consulting"... I want to run NixOS in prod, and having clear options for "oh crap we need help" is actually a big deal.
<danderson> I don't think we'll need it right away, but it's another "confidence signal"
<gchristensen> ^
<danderson> vs. "oh we can hire thousands of people who can consult for Ubuntu"
<samueldr> and psst, the website already tries to do that
<garbas> drakonis: to be honest 3 years, but never had the time to step up and do it. i finally came around and started talking about it.
<genesis> i'm no "lol consulting" i was just reading an how to make insider job
<gchristensen> domenkozar[m]: I don't know where you see it is about promoting consulting?
<garbas> cole-h: danderson: i'll try to record it, if everybody will agree and if i managed to firgure it out how to do it. :)
<genesis> all requirement are made to make only people inside the already known group of people who decide, let decice how to promoting nixos/nix .
<drakonis> ah, that's a lot of time
<danderson> garbas: sounds great :)
<domenkozar[m]> I think we should move towards "you don't need a consultant to use Nix", which is the opposite of what the responsibilities of the marketing team says
<drakonis> its a excellent move.
<genesis> so to participate, be "connect with companies that offer Nix/NixOS consulting"
<domenkozar[m]> so it's quite an opposite goal
<FRidh> aside from, have a page showing what organizations are using Nix in production (of course they need to be open to sharing that)
<cole-h> garbas: What danderson said -- works for me :P
<drakonis> now, looks very good.
<danderson> you can do both, IMO. Make Nix easy to use and operate, and also offer paid services to help people with harder problems.
<drakonis> it'd probably serve as a good starting point for attracting companies
<danderson> making it easy to find consulting services doesn't necessarily mean Nix can't get easier to use.
<danderson> (and in fact, a good consulting service *wants* to make things easier to use, because it lets them move "up the stack" and focus on solving bigger, harder problems, instead of selling "help how do I nix" a hundred times)
bhipple has joined #nixos-dev
<colemickens> builtwithnix is not very information dense and gives off a generic vibe that I can't describe that I don't like. So much scrolling.
<garbas> domenkozar[m]: i think i talked to you already and explain you what this marketing team is all about. it is to bring nix mainstream, or at least push in this direction. can this be first step: no. would we want it? ofcourse.
<MichaelRaskin> I guess «bringing to mainstream» requires covering mainstream, the companies who just fail at running IT projects in-house.
<drakonis> colemickens: it has a flair i can't deny
<garbas> domenkozar[m]: picking on exact wording of that announcement is exactly what i hope to avoid. one way is to ask if your idea is also part of what is behind marketing team. i'd hapily explain.
<colemickens> I like the flair, but I think there's a healthy compromise between the clean FOSS simplicity of the current nixos site and something that feels like a startups landing page
genesis was banned on #nixos-dev by gchristensen [*!*@unaffiliated/genesis]
genesis has left #nixos-dev [requested by gchristensen (we don't talk about contributors that way.)]
<gchristensen> for context, the removal of genesis is temporary, and in response to their comments about a dearly appreciated contributor.
<danderson> It's definitely hard to balance OSS projects between "appealing to companies" and "friendly to OSS"
<danderson> Anyway, I'll step away now, I don't have any marketing skills that would help. But I'm looking forward to seeing what y'all come up with :)
<drakonis> to be fair
<drakonis> there might be ways to help here that aren't going to require marketing skills.
<samueldr> I've seen needing consulting as being a need already, that wouldn't be filled by "making it not required"
<samueldr> it's not an either/or, it's an and/plus
<garbas> danderson: it is hard, but i think few projects did it in the past. i think rust page is a nice example of a balance., but that might be only my perspective
<domenkozar[m]> garbas: right, but there's a tone you need to decide on and the audience
<danderson> drakonis: happy to chip in as I'm able to, then. Right now I'm trying to contribute by reducing the security bug backlog, which is more directly useful to my "nixos in prod" plans.
<gchristensen> danderson: thank you for that, that has been a long-term interest of mine, as well
<gchristensen> danderson: if you need a big machine for building things for it, I can provide access to one
<garbas> drakonis: +1 i'm new to "marketing" too. but i'm ok to learn if everybody else is
<danderson> gchristensen: I have a bunch of large machines in my basement, but honestly so far I've just been doing some basic checking and then letting ofborg figure out if I broke the world
<bhipple> gchristensen: speaking of big machines, is there a way to connect a temporary OfBorg CI bot to the cluster? Or in general how hard is it to add one?
<gchristensen> bhipple: right now, Difficult
<danderson> not ideal... But that's what I can do until I build a bigger machine :)
<bhipple> I tried my hand at reading the instructions on, but I noticed a lot of things were out of date
<gchristensen> the idea of it to start with was that it would be exactly that -- a collection of people offering CPU time. but I don't think builders are the bottleneck?
<garbas> domenkozar[m]: agree. but we can at least start the conversation now and bring it out and test it/listen to the community... we don't have to decide everything in a week.
<garbas> domenkozar[m]: we can do otherthings meanwhile while also talking about "the big picture"
<drakonis> the first thing we could do is start with a main page redesign, as you said earlier, rust's page looks good.
<bhipple> Under normal circumstances I don't think builders are the bottleneck necessarily, but they were for some of the PRs I was sending. Some details in
<danderson> gchristensen: anyway, if it turns out I'm breaking stuff too much, I might ask for access to a big machine. So far I'm betting on the fact that I'm not making *very* scary changes.
<danderson> thanks for the offer!
<drakonis> it has a good tagline
<bhipple> E.g., if I know I'm about to send 20 compile-heavy PRs and I have a lot of build capacity, it might be nice to temporarily connect some of my capacity to help out. The job is now done, but once I implemented the rate limiting the bot spent the vast majority of its time just sleeping for OfBorg to catch up :D
<gchristensen> great! it is hard to do too much damage, since everything is pretty well checked before it goes out
<drakonis> its a good rust in a nutshell idea
<drakonis> s/idea/tagline
<bhipple> This is a relatively low priority thing, tho, since I think my use case was rare (even for myself). BTW as part of that we now have a *ton* of improvements to Ryan's nixpkgs-update bot :)
<drakonis> NixOS would be be well described as "A modern way to compose your systems", Nix would be "Reliably build your software"?
<samueldr> y'all know the meeting isn't today right now?
<drakonis> yep?
<samueldr> just checking
<gchristensen> haha
<gchristensen> <3 y'all
<MichaelRaskin> I guess people decided meeting will be hard and video
<MichaelRaskin> Should prepare in a more friendly medium
<drakonis> got a cool 4 days until the meeting takes place
<drakonis> a whole lotta time
<MichaelRaskin> BTW I think I have seen quite a few projects who gave up and have separate companies-facing and OSS-facing websites
<drakonis> debian does that
<danderson> There's definitely two audiences, and right now I'm half of each: nerds who want to play with the cool tech, and business people who want to do business things
<domenkozar[m]> yes, there needs to be an user and developer audience and right now it's both
<danderson> But I think it's possible to have a marketing front page that works well for both, and then funnels them into different places for what they're trying to do
<danderson> i.e. have a pretty and polished front page, with 2 ways out: are you looking for business whitepapers/consulting/etc., or are you looking for the nixos manual?
<danderson> that way you can still point everyone at a single website, and they'll find their own way from there. Nicer than separate sites where you have to figure out ahead of time what person X wants
<drakonis> that sounds good.
__monty__ has quit [Quit: leaving]
<gchristensen> this is a bit off the topic, but if anybody sees disrespectful GitHub activity from "bignaux" -- please let me know?
<cole-h> I'm unfamiliar with automake's Makefiles -- are variables like `$(sysconfdir)` inherited, or something?
<cole-h> (well, Makefiles in general really)
<cole-h> I ask because the zfs pkg uses `substituteInPlace` on only 2 files reference `$(sysconfdir)`, but the tree has multiple other files.
<cole-h> I'm really just wondering why "every" file doesn't get `substituteInPlace` run on it to replace `$(sysconfdir)`
<gchristensen> I think it is inherited yeah
<cole-h> Whatever. It obviously doesn't matter that much, or somebody would have filed an issue :P
FRidh has quit [Quit: Konversation terminated!]
abathur has quit [Ping timeout: 255 seconds]
<gchristensen> q/topic
<gchristensen> q/topic
<gchristensen> hi yes I computer good
<samueldr> you know, q and / are at the opposite end of the keyboard
<samueldr> :3
<drakonis> what if... he's using dvorak
<samueldr> he is, that's exactly why I'm saying that
<samueldr> and anyway it's a falsehood, q and / are two keys sidewards, and one key up/downward from each other, though one is shifted ;)
<drakonis> haw, the key distance between q and / is the same as querty
<Profpatsch> ofborg@home
<adisbladis> <3 Profpatsch
<{^_^}> Profpatsch's karma got increased to 12
<Profpatsch> you know what time it is?
<Profpatsch> it’s time to RIIR
<drakonis> RIIR?
<drakonis> rewrite it in rust...
<drakonis> its always time to RIIR
<adisbladis> aanderse: You around?
<aanderse> yes, until i have to put the kids to bed at which point i'll be unavailable for a few minutes
<aanderse> whats up?
<adisbladis> sshd[817]: error: AuthorizedKeysCommand path is not absolute
<{^_^}> #82413 (by aanderse, 2 days ago, merged): nixos/sshd: add authorizedKeysCommand and authorizedKeysCommandUser options
<adisbladis> We need a check for when it's `none` and omit it
<aanderse> :\
<aanderse> none is the default and i checked that...
<adisbladis> I'm fine with fixing it since I'm touching the SSH module anyway
<aanderse> i... don't know what to say
<aanderse> i specifically tested the case where i didn't modify it
<aanderse> :\
<aanderse> ie. it was "none"
<aanderse> well... i apologize for the hassle
<adisbladis> It's alright :>
<aanderse> i was actually just putting in another ssh change
<aanderse> i'll take care of that
<aanderse> i'm in here
<adisbladis> More of a fyi than shame on you :>
<adisbladis> aanderse: What are you doing?
<{^_^}> #82620 (by aanderse, 10 seconds ago, open): nixos/ssh: silence ssh-keygen during configuration validation
<aanderse> i finally realized where those random key arts were coming from
<adisbladis> Ok, good
<aanderse> :p
<adisbladis> Just wanted to make sure we're not stepping on each others work :)
<adisbladis> I'm making a test for
<{^_^}> #81661 (by adisbladis, 1 week ago, open): pam_ssh_agent_auth: Honour services.openssh.authorizedKeysFiles
<aanderse> as someone who has a locked down authorizedKeysFiles value set across my nixops network i approve of that PR 👍️
<aanderse> heh
<adisbladis> =)
<aanderse> adisbladis: unfortunately i'm not in a position to test this atm... any chance you can give this a quick test and review?
<{^_^}> #82621 (by aanderse, 44 seconds ago, open): nixos/sshd: only include AuthorizedKeysCommand and AuthorizedKeysComm…
ixxie has quit [Ping timeout: 240 seconds]
<adisbladis> aanderse: LGTM
<adisbladis> I'll do a proper test and merge in a bit
<aanderse> thanks, and again apologies
<aanderse> i'm trying to recall how this went wrong still :\
<adisbladis> aanderse: Iirc it's not a critical error
<adisbladis> So it's mostly polluting logs
<aanderse> oh... i thought you meant it stopped openssh from starting
<aanderse> it is possible my logs weren't verbose enough
<aanderse> ahhhh
<aanderse> ok